Committee on Culture and Education Ordinary meeting
Legend
Transcript markers
- text Sentence being played
- text Search result
- text Intervention cited by the link you opened
Click a sentence to play the video from that moment. Select text to quote it with the speaker, timestamp and link.
More options
CULT European Parliament
Follow its sittingsSpeaker 4
On the other hand, consequentialism is about the outcomes or the consequences of an action taking precedence over rights or obligations. So there, acting ethically means undertaking actions that result in the best overall result, the greatest good for the greatest number of people. Now the reality is that a lot of contemporary ethical frameworks, they tend to rely on a blend of these two approaches without really understanding either. And in particular, they also ignore more recent approaches, such as the ethics of care. And in particular, they ignore the fact that Europe is a pluralistic tapestry of multiple faiths and secular convictions.
So all of those I've talked about emerged from the Judaic-Christian tradition, and we should also be looking elsewhere to make sure that we are more inclusive. So the ethics of AI, Ed, to begin with, it needs to build upon the ethics of AI. You know, every stage of the AI development pipeline raises different ethical issues, such as how data is gathered, how data is trained, the ownership of the tools, the outcomes. In 2018, the European Commission's high-level expert group on AI published the landmark ethics guidelines for trustworthy AI. This identifies a wide range of issues, including bias, fairness, and also the ghost workers of AI.
In other words, it identifies and focuses on what we might call the generic issues of AI, but it doesn't look at the domain specific ethics relevant to the domains in which AI is applied such as health and transport or education. Now while its guidance is based on rights enshrined in the charter of fundamental rights of the human European Union, which is a deontological lens, it's actually unclear how such principles might be actioned in a specific context. And there is a growing body of work, with some exceptions, on the ethics of AI, which doesn't address education-specific issues. And the exceptions are work I've been involved in such as the book from 2020, The Ethics of AI in Education, Practices, Challenges, and Debates, but that kind of thing is one of the few. A more typical example is the European Commission's ethical guidance on the use of AI and data in teaching and learning for educators from 2022, which again provides really useful guidance.
But, again, it mostly focuses on the generic AI issues. The OECD's opportunities, guidelines, and guardrails for effective and equitable use of AI in education, that take a similar approach. So they don't address the specific ethical issues of AI in education. They don't talk about the purpose and aims of education. What constitutes use for knowledge?
The choice of pedagogy? The impact of AI on human development and the lack of robust evidence for effectiveness or safety. We often hear that it brings lots of opportunities, but the reality is we don't have any real evidence that those opportunities are really coming to bear. We have no robust independent evidence of the scale for either the effectiveness or the safety. These tools, these policies, though, they also tend to anthropomorphize AI, which leads people to assuming that AI has more capabilities than it does and has actually contributed to the deaths of young people around the world.
And existing ethical frameworks don't address the insufficient debt, the impact of AI systems and AI companies on teachers and learners. For example, on teacher and student agency. It doesn't address issues such as automated abdication where teachers hand over their task to AI and lose control. It doesn't talk about cognitive atrophy, where learners are benefiting from using AI only when they are actually using it. As soon as it's taken away, it's lost. The learning is not happening.
And as I say, it doesn't address the fact that there's no robust independent evidence of scale for the effectiveness or the safety of AI or AIED, which means that every time AI is being used in education today, we are effectively experimenting on children outside of any ethical controls. Now we hear a lot about the digital divide, and the digital divide is getting more complex. So AI is exacerbating the digital divide between those who have access to the technology and those who do not. But, actually, the digital divide is beginning to flip, and what we're seeing is that those from wealthier backgrounds are continuing to have access to human teachers, whereas a lot of young people from disadvantaged or challenging backgrounds are increasingly having to make do with computers. We also know that the AI Ed risks violating the UN Convention on the Rights of the Child because it reduces students to data points, because it is eroding autonomy, and because it is reinforcing biases.
And in particular, it's facilitating commercial exploitation, treating children's interactions as unpaid labour to train proprietary models for profit. Now one of the ways in which we might help improve the ethics of AI Ed is to support all learners and their teachers to become genuinely AI literate. Now what AI literate means is very different. We've done a review. We've got 115 different frameworks of AI literacy, and they're all different.
But, actually, most of them, such as the EU and OECD's AI Lit that was published last year, focus on teaching young people how to use AI. But in my opinion and the opinion of my colleagues, AI literacy should not be about training people to use AI. It's an uncertain technology. But instead, it should focus on helping people to thrive and flourish in a world increasingly dominated by AI. So AI literacy should focus on the ethical risks that emerge at every stage of the AI development pipeline, such as the impact on the environment, the backdoor commercialization of education, issues of surveillance, and algorithmic determinism.
But finally, even if the ethics is properly grounded in moral philosophy, even if the ethics addresses both the generic AI issues and the specific domain-specific AI ed issues, nothing will happen without enforceable governance mechanisms being in place. Mechanisms such as mandatory auditing, risk assessments, and specific legislation to protect learners and their teachers. The important thing will be to shift the ethical burden from users, from teachers and students, to the developers and the purveyors of these tools while protecting teacher autonomy and student agency. So in conclusion, the ethics of AI Ed is complex, but it's fundamentally important to the future of young people, their teachers, and education more broadly. If the ethics of AI Ed is to be successful, it must be properly grounded, grounded in moral philosophy, and grounded in the specific issues of education.
And it must be supported with AI literacy that is designed to help young people and their teachers thrive and flourish in a world increasingly impacted by AI. And it must be supported with robust and enforceable governance mechanisms. Thank you very much for listening.
Thank you.
Speaker 5
For your presentation, professor. And now, Professor Carisi, the floor is yours.
Speaker 6
Good afternoon. My slides are not there yet. Yeah. Good afternoon, everyone, and thank you for the opportunity to talk today. My talk will focus on the pedagogical dimensions of artificial intelligence in education.
I won't describe what AI can do. Rather, I'm going to elaborate on how AI transforms and reshapes teaching and learning. As AI becomes increasingly embedded in education, the central question is no longer if we use AI in education, but how we ensure that it supports meaningful learning, human development, and democratic values. Artificial intelligence is already transforming multiple domains, from science and health care to society at large. Education is part of this transformation.
However, in education, this transformation is not only something technical, but it reshapes how people learn and actually shapes the brains of our future societies. AI changes how knowledge is formulated, how students interact with the knowledge, and how teachers design their learning environments. This raises a critical question for policymakers. Are we designing AI tools according to our vision for future societies, or are we only allowing AI technology to shape education? Present recent evidence presents a nuance and somehow concerning result.
On the one hand, when we introduce AI systems, we see that they improve student performance during practice. On the other hand, studies show that when AI supported activities, when we remove AI, students' independent learning and performance decrease. So we must distinguish between performance and learning. From a policy perspective, this distinction is essential. Education is not about short-term optimization of answers, but about development of durable understanding, critical thinking, and autonomy.
So in this talk, I will briefly address five points. What do we mean by pedagogy in the era of AI, the opportunities and challenges of AI in classrooms, the education goals and how we could and should prioritize, the role of teachers and school leaders and educational stakeholders, and finally, the key policy considerations? Pedagogy often is misunderstood as simply a method that teachers use in their classrooms. However, in reality, pedagogy is much more complex. It includes education, the definition of educational goals, teaching approaches, the evaluation of methods, the orchestration of social interactions inside the classrooms, the tools the students use, etcetera.
AI introduces a wide range of tools into this ecosystem, like intelligent tutoring systems, adaptive learning platforms, chatbots, etc. But at the same time, we know that these tools are not neutral. They embed assumptions about what learning is, what should be measured or could be measured, and how students interact and behave. Ideally, pedagogical principles should guide and interact with the design of these AI tools. However, we observe that currently, this interaction is mon there is a unidirectional interaction.
So the tools design and affect the pedagogical goals and the methods. We acknowledge that AI offers significant opportunities for learners. It can help scale access to effective pedagogical practices, support teachers in their daily work, enable personalized formative assessment, etcetera. However, these opportunities come with certain important risks. We observe a growing misalignment between technological design and pedagogical goals.
Most importantly, we observe that there is an increasing commercial influence within these educational ecosystems that may introduce certain business models and reduce pedagogical autonomy. As a conclusion here, it's not we want to emphasize that we don't want to say if AI is inherently beneficial or harmful. The conclusion here is that their impact depends on how it is designed, governed, and integrated into educational practices. If AI is transforming society, then education must prepare students for this transformation and for future societies. These require an examination, a re-examination of educational goals.
We needed to place greater emphasis on critical thinking and epistemic judgment. AI literacy does not only include technical understanding, but also awareness of its social implications. Ethical and reasoning, ethical reasoning and value-based citizenship is a priority. And at the same time, we must ensure that learning environments remain inclusive and responsive to diverse learners. As such, education should equip students with the capacities to understand, question, and shape AI systems responsibly.
In this context, the role of teachers, educational leaders, and education stakeholders is central. Teachers are the designers of learning environments and mediators of knowledge. Their role is to ensure that AI supports meaningful learning processes, and that they inspire students and motivate their learning. Evaluation methods consider advances in AI tools, of course, but at the same time, they should be defined by the educational goals. Effective integration of AI also requires specific pedagogical approaches.
One example is constructionism. Involve this, this moves students from simply consuming AI outputs to actively imagine, reimagine, building, experimenting with AI. Another example is collaborative learning. A third is value-based pedagogy. All these are not new approaches.
They are approaches that have been here for many years, but they support what we call epistemic knowledge and metacognition. Students become active actors of their own learning and evaluators of their own learning. And at the same time, we know from research that this is a way to motivate students. At the same time, assessment remains a critical aspect. This brings us to a key policy question.
Who defines the pedagogical principles embedded in AI tools? How are these tools evaluated before they are adopted in schools? And how can teachers maintain authority over learning processes in AI-mediated classrooms? The European Union already has a strong policy framework, including the AI Act, the Digital Services Act, and the Digital Education Action Plan. The challenge now is to ensure that these frameworks are implemented in ways that place learners at the centre.
Looking ahead, three elements are essential for the responsible integration of AI in education. Trust, multistakeholder collaboration, and accountability. This requires clear definition of roles and responsibilities among actors, and transparent evaluation mechanisms. And here we show an example of a framework for the facilitation of transparent multistakeholder collaboration in the form of a sandbox. Let me conclude with three key messages.
First, learners' well-being and educational goals, and not technical capabilities of AI systems, must guide the design, development, and use of AI in classrooms. Second, teachers, school leaders, and educational stakeholders are central to this process, should be supported, and students must remain the core of AI in education. Third, trust can only be achieved through transparency and clear accountability among all actors, including industry, governments, researchers, and most importantly, parents and students. Artificial intelligence has the potential to transform education, but transformation alone does not guarantee progress. Progress depends on whether we can align AI with long-term learning goals for students and democratic values.
Thank you very much, and I'm looking forward to your questions.
Nela Riehl
Thank you for this presentation. And now I pass the floor to professor Gunther, who is joining us online from Japan.
Speaker 7
Good afternoon. My name is Irene Angelica Gunther. I'm a professor of computer science at the University of Duisburg-Essen, and I work on artificial intelligence in education. Next slide, please. If you see my slides.
Not yet.
Speaker 7
Today, I will discuss the main trends, risks, and policy recommendations regarding the use of artificial intelligence, with a focus on cognitive aspects and compulsory education. Next slide. Young people are the largest user group of AI tools across all the member states. 64% of them used generative AI in 2022, and out of those, 40% used AI tools specifically for formal education. What is significant is that their primary motivation for using AI is efficiency and convenience, so saving time, making tasks easier, for example, obtaining complete answers to learning tasks.
From a learning science perspective, this is the pattern most associated with cognitive harm. Next slide. Now let's establish why children are a specifically vulnerable group. Following the theory of cognitive development, children progress through four sequential stages. So as infants, we start learning through direct physical interaction with the environment, through our senses.
Then preschool children develop language and learn associating words with concepts, progressing to thinking logically about concrete objects and events during primary school. And from 12 years old onwards, they develop abstract, hypothetical, and systematic reasoning. Now what matters here is that these capacities most at risk from problematic AI use, logical reasoning, abstract thought, self-monitoring, are still being constructed during the school years. Another important implication is that a seven-year-old, a 12-year-old, and a 16-year-old, they all face different risks from the use of AI, and this is why an age-differentiated governance is essential. Next slide, please.
From a cognitive science perspective, learning is defined as a change in long-term memory. So what happens is information enters the human brain using our sensors, our ears, our eyes, and moves through the working memory, which is limited in capacity. It gets processed, and then it is transferred and encoded into long term memory, forming connections with existing knowledge, so that we can retrieve it later when we need it. The critical point here is that, in order for this encoding in the long term memory, so in order for learning to happen, it is necessary that we put some cognitive effort there. So that explains why tasks that feel difficult in the moment that we practice them produce stronger long term retention.
So they stick better in our brain. Next slide. The cognitive dimensions that we will discuss today are organized around four core processes: attention, memory consolidation, reasoning, epistemic judgment, and metacognitive regulation. Next slide. And what I will not talk about today is digital literacy, and that is because a student can be digitally literate.
So a student can be fully competent with AI tools. They might be fully aware of the risks and limitations, and still experience cognitive harm. Next slide, please. And this is important to point out because it means that AI literacy education, while necessary, is not sufficient as a policy response. Next slide, please.
I want here to make another distinction between general purpose AI and educational AI. So general purpose AI, for example, large language models, is designed for broad use across domains and audiences. It is not optimized for learning, and it responds to prompts without any regard for the learner's cognitive or knowledge state. On the other hand, AI-enhanced educational technologies, so the field that I called AI ed, are purpose-built within a learning science framework. They are designed in a way to support students' cognitive development.
The cognitive risks I'm about to describe are most strongly associated with general purpose AI when used in an educational context without pedagogical mediation. Next slide, please. I will now turn to the four cognitive risks that have been identified so far, and I will take each briefly in turn. Next slide, please. The first risk concerns attention and memory consolidation.
When we use AI to generate an output, practically what we do is bypass the process by which knowledge is constructed. So when a student uses an AI model to receive the finished product, a summary, an essay, a solution, they have not performed the cognitive work through which information is encoded in their long-term memory. And the result of that is shallow processing, and consequently weak encoding. A key concept here is the expertise reversal effect. What works as cognitive offloading for an expert?
So what an expert does when using AI is to handle routine tasks so that they can focus their attention on other problems. It does not work like that for a novice. Novice learners lack the foundational knowledge that makes offloading productive. When novice learners use AI to do the cognitive work, they are not freeing up capacity for deeper processing. They do not offload the cognitive work.
What they do is they outsource cognitive work. So instead of doing the thinking in their brain as they should, an external agent is doing it for them. Next slide, please. The second risk is dependency, and we know from research on automation bias that humans systematically overtrust automated outputs, especially when they are under pressure and including when the automation is wrong. For children whose critical evaluation capacity is still forming, this habitual reliance on AI can create strong dependency that can become self-reinforcing, so more likely to happen again and again, and it even extends beyond schoolwork.
Next slide. The third risk concerns the development of fundamental skills such as reading and writing, and these are skills that we acquire through practice. When AI tools do perform these skills on behalf of the learners, practically, they remove the practice that builds them. Like the colleagues already mentioned, the paradox here is that the short-term output quality may improve. However, the underlying skill weakens.
So, we've seen from recent studies that students using general purpose AI tools perform better in the short term. So they might get a better grade. But when they get another test and the AI access is removed, they perform worse. Now for compulsory education, where the objective is to develop that kind of foundational skills, we have a structural problem. The metric that we most easily observe, so the task quality, is decoupled from the outcome that we actually care about, and this is skill acquisition.
Next slide, please. The fourth risk is perhaps the most consequential for long-term development, and this is the erosion of metacognition and epistemic autonomy. So metacognition, the capacity to monitor, evaluate, regulate one's thinking, is built also through effortful practice. And when AI performs these functions, then the practice that builds metacognitive capacity is removed. And this is precisely captured by the term metacognitive laziness.
The student who lets AI evaluate the work for them to flag the errors, to suggest revisions, is not exercising the self monitoring capacities that constitute independent thinking. And an even deeper concern here is epistemic delegation, so the transfer of epistemic authority from the learner to the system. A student who accepts AI generated judge judgment not just fails to practice reasoning. They are actively learning to defer it. Next slide, please.
So taken together, these four risks, the briefing argues that children in compulsory education are a specifically vulnerable group and that primary age children are the most vulnerable of all. And that is because working memory and executive function are still maturing through adolescence. Critical thinking and epistemic reasoning develop through practice during school years, and they are constructed through experience. And capacities that are not exercised during the formative school years may not develop fully. To put it in simple words, a higher education student who delegates cognitive work to AI may simply be underperforming. However, a child who does so consistently during primary school might be failing to acquire the cognitive architecture that is necessary for subsequent learning.
Next slide. The existing regulatory frameworks. So the AI act, the GDPR, the Council of Europe framework convention, they establish requirements around transparency, human oversight, risk management, and these are significant achievements. However, they do not address the cognitive impact on child users. For example, the AAAC classifies AI systems in educational settings as high risk under Annex 3.
However, the classification framework does not currently require any demonstration that an educational AI system supports learning. An AI system that increases short-term performance metrics while undermining long-term skill development most likely would pass a regulatory review right now. Next slide. To address this, the briefing proposes three concrete measures. First, establishing a cognitive impact assessment as an education-specific layer within or alongside the fundamental rights impact assessment under article 27 of the AI act.
This assessment would examine cognitive load, attention effects, developmental appropriateness, risk of creating dependency, and so on. Second, introducing minimum evidence standards for pedagogical efficacy claims as a condition for AI systems used in compulsory education. This in combination with a mandatory disclosure register analogous to the EU database for high risk AI systems under article 71, which providers should publish evidence based for their claims. And third, requiring pedagogical evaluation before public school procurement as a condition of award. So something analogous to what we have now, regarding accessibility requirements for digital tools.
And this would create an EU-wide baseline building on national frameworks that are already developing or have developed in several member states. Next slide, please. So the policy challenge is not prohibition. It is designed. The question is not whether to allow AI in education, but how to ensure that the conditions for cognitively safe and productive AI use are systematically built into how tools are developed, procured, deployed, overseen.
Overregulation is necessary, but it's not sufficient. Curriculum and teaching practices must also adapt, and that's why investment in teaching preparation is as important as investment in regulatory architecture. Next slide. Let me close with a summary of where I believe we are. The widespread adoption of AI exposes young learners to real cognitive risks.
However, this is not an argument against AI in education. We know from decades of research in AI in education that purpose-built educational AI deployed with proper pedagogical scaffolding with teacher mediation can genuinely support learning. So the challenge is to ensure that these conditions are the norm, and not the exception. Next slide. The AI act and European educational policies are still evolving, and I see this as a window of opportunity.
And with this message, I would like to thank you. Next slide. The full briefing paper, I believe, is already available to you. I'm happy to take questions, and, of course, I'm available for any follow-up the committee may wish to pursue. Thank you very much.
Nela Riehl
Thank you from our side, too. And before we move on to the coordinators and the catch of the eye, I'm afraid I will have to take my leave of you today, but Mr. Sorensky is very kindly taking over the work of sharing the meeting. So I hope you have a wonderful afternoon, and I thank you very much for being here today.
Speaker 8
Thank you, chairman. I'll be speaking in Italian. My institutional view is accompanied by my direct experience in the school world, first as a teacher and then as a director of the school. AI in classrooms, as has been said, and I would like to congratulate the speakers, represents a huge opportunity not just in the service of the teachers, but also if it helps to develop the students as well. As a teacher, I saw how this is a key with capability of interpreting the needs of students and stimulate their critical thinking, their reasonings, and their creativity.
As a director of a school, I was able to see how organizational choices and school policies can support or limit educational innovation. AI, as has been said, is capable of reinforcing the collaborative learning experience, in particular targeted feedback, differentiating the precourses. Under it obviously has to be under the guidance of the teacher. But before that, the teachers need to be trained, involved in the choices, and responsible for the educational pathway of their students. Together with ongoing training and practice, we need to put the teachers at the heart of this, ensuring that the support instruments for professionalism are there.
So I'd like to ask, how can we guarantee that AI really reinforces learning and the key role of the teachers without compromising their role. Can teachers really use AI in the classroom to stimulate critical thinking and still maintain an active role in learning? Thank you.
Thank.
Speaker 9
You, and thank you very much to the experts for the great quantity of information that, first of all, we need time to organize. So it is the basis, cognitive skills are the basis of our education and of science. Young people need to be given to understand this by the teachers. So they need to know what their opinions are and what information they are confronted with. So my hope is I'm quite hopeful because young people in the European Union, according to surveys, are clear about the fact that they don't have enough digital competences and digital skills.
Is this an indication that there is a will there? How far do young people really trust materials which are produced by artificial intelligence? And I'd like to ask Professor Holmes. He said there was only one generative AI, which is in use in the educational system. And Professor Carys, we heard that generative AI could achieve good results in mathematics.
And so for me, I'd like to know how specific and concrete the examples are that lead you to say this. And when people are learning languages, where is it critical? So we hear that there won't be a need for interpreters. But I don't agree with that because every expression that somebody used can have many different meanings. And here in the European Parliament, I don't think that artificial intelligence can be successful.
But when we are talking about project work as well, when is it useful to use AI and when isn't it? So, and how can teachers see that AI has been used for students' work? And to achieve this cognitive work with young people or with adults, is it possible to prevent the use of computers and tablets for certain learning targets or for testing of learning targets?
Speaker 10
Thank you, Chair, for giving me the floor. This is an extremely interesting hearing, and we hear about these aspects, both cognitive and ethical. I have asked myself the same questions we heard here. However, I'm concerned to hear that people who know more than I do and we do also have questions, even more questions than I have. I have some questions for Professor Holmes.
Speaker 18
My question is, do we have real control over the data gathered on pupils? Who is gathering this data? And is there a risk of profiling people starting from a very early age? And through this profiling, someone would know who knows what, who does what. And so these profiling will lead to some targeted approaches to people.
Speaker 13
Then maybe governments, national governments should introduce some specific norms, some specific rules that would regulate the use of artificial intelligence in schools. To professor Harizi, I have some questions for you too. How can we understand where is the dividing line and who is to draw it between the artificial intelligence helping a teacher and the artificial intelligence replacing the teacher? And what is the role of the legislator here? Next question.
Speaker 10
Is there a risk of teachers losing their authority, their credibility if it is the AI which assesses, evaluates the pupils? To professor Kunta, do we have statistical data which we could use as evidence that the frequent use of AI reduces the capacity to think independently? And to all of you, how could we avoid creating generations that rely more on answers rather than on asking questions? And how could we optimize education thanks to AI, but also at the same time, we will develop the capacity for critical thinking of pupils. After all, these are future citizens, future European citizens who are supposed to make their own independent decisions.
Speaker 12
So how could we preserve the human in human beings? This is a philosophical question. Thank you.
Speaker 16
Thank you, chairman, and thank you to our speakers. Artificial intelligence is ever more present in our schools. It can offer significant opportunities for learning, but it also gives rise to some very serious questions about the role of teachers, the quality of education, but also the development of critical thinking in the students. Artificial intelligence should not replace the educational relationship, nor transform learning into an automatic procedure. So because of that, I'd like to ask, which specific instruments would you make available to teachers for using AI in a responsible and effective way?
And how can we guarantee that teachers are adequately trained to be able to integrate this technology without losing the human values of education? Thank you.
Speaker 15
Yes. Thank you very much, chair, and thank you very much to the distinguished guests and experts who have shared their experience and knowledge about AI in education with us. It is true that artificial intelligence represents a radical change in education and learning for teachers. It could serve as a tool which would help them to better organize the subject matter, the curricula, and monitor the progress that their students are making, and obviously adapt classes based on the abilities of their pupils and students. And for the pupils and students, it also represents a way of having better access to knowledge and to be able to learn at their own pace.
Speaker 20
I come from a country, from Greece, where our government is. But they wish to link technology with education. But, obviously, this has to be done in a very specific way, in such a way that artificial intelligence does not turn from a useful tool into a threat for both students and teachers. So I do have a series of questions. And any of you, any member of the panel could reply to this. How can we make most of the benefits, the advantages of AI without this posing a threat to young people?
Speaker 17
And obviously, we want to make sure that critical thinking is developed. And we have to make sure that AI is of benefit to this support to students, and does not deter from critical thinking. And how can AI help teachers? And how can it help, or how can teachers make sure that their students are not using AI to do their homework as it were, and replacing the teacher by AI by having a kind of twenty-four-hour teacher and lecturer. So whether we're talking about schools or universities.
Thank.
Speaker 19
thank you very much. So we've had a first round of questions, and I'd like to turn to the guests, the members of the panel, and ask them to reply. So perhaps if you could reply in the same order that you took the floor, reply to your questions. Professor Holmes, please.
Speaker 3
And thank you very much. I mean, a huge number of questions there. I really wanna start with this question of this opportunity. Every time I use one of these tools, and for my work, I play with them a lot, I'm constantly amazed at what they can achieve. What the computer scientists have achieved with this is extraordinary.
Speaker 4
But within a very short time, I'm incredibly disappointed because these tools are not good enough for education. And we hear time and time again that these tools, and so my colleagues more or less said this, that these tools bring wonderful opportunities to education. It's just there are some problems around the edges. We need to deal with the problems around the edges, and then we'll be fine. But I repeat, there is no independent evidence at scale that any of these tools used in an educational context are either effective or, more importantly, safe.
So in other words, every time one of these tools is being used in a school without any regulation, because there isn't any regulation that's relevant, we are experimenting on those young people. That is the truth. If you wanna experiment, that's your prerogative, but that's what we're doing. So when we're tasked, where is the evidence? I often hear this phrase, AI brings amazing opportunities for education, but it also brings risks and challenges.
Well, I and my colleagues can give you lots of evidence about the risks and challenges, but you can't give me good, independent quality evidence that these tools are effective and safe for our young people. If you can, give it to me. I'd love to see it. And what we do know is that over recent years, scientists have become less confident about these tools. We know that the industry is becoming less confident in these tools.
We know that generative AI, which is what most people mean when they talk about AI, is in danger of bursting its bubble, and we don't know what's gonna happen there. So it's a very complex, fluid situation. But one of the questions we were asked is, you know, how do we get young people to trust these tools? How do we get teachers to trust these tools? Well, the answer to that is simple.
We need the tools to be trustworthy, and the problem is they're not trustworthy. They generate nonsense. They're prediction machines. They look incredible, but when you dig below the surface, they generate nonsense. They generate biases.
They generate fabrications. And we really need to get under the heart of this to make sure, that we understand what these tools actually do and not what just they appear to do. The difference between the appearance and the reality is immense. And to support our students and our teachers and education more broadly, we need to really understand this far more. Now, colleague at the front, my sincere apologies if I misled you.
I didn't mean to say that there was only one Gen AI-type tool. What I meant was of the 25 types of AI being used in education, one of them was Gen AI. But what has happened more recently is that Gen AI is now embedded in many of the other 24. So it's becoming increasingly complicated. We talk about these tools undermining cognitive development, and that's where we are seeing the evidence of what these tools are doing.
We talk about how we support teachers to use these tools to integrate them into their education. Well, I don't want them to integrate into their education because I don't believe that this is positive for our young people, is positive for our children, is positive for our society going forward. Now in 20 years' time, I have no idea what the AI tools will be able to do. Invite me back in 20 years, and maybe I'll change my tune. But today, based on what we have, based on the technology that's out there, we don't have this evidence.
And what we are doing, particularly in Europe, we're giving lots of evidence and data to our American colleagues who are using it in ways that are really quite problematic. Now we do have smaller devices here, I know, but the vast majority is still going out. And it's undermining the relationship between teacher and the student in the classroom because the teacher, when they use these tools, has no control. They have the only control they have is, do I allow the child to use the tool in my classroom? That's it.
After that, it's entirely in the control of the tool, but actually, it's not. It's in control of the companies behind the tool, which includes some really quite despicable organizations and people, again, typically in the USA. And one of the questions was, how do we make sure that young people are not using these tools? Well, give up. You can't.
Right? It's not possible. There is no tool that can tell you definitively that that's not happening. But what we are seeing, one of the things I think is interesting, you go back five years and speak to your average students, and they were stressed people. By definition, students are stressed.
They're stressed about their classes, about their exams, about their nightlife, about their food, about their home. Since their arrival of ChatGPT and the others, we were told these tools were going to help the students. But, actually, today, they've got all of those stresses that they used to have. But in addition, they're now stressed about generative AI. If they don't use it, will they fall behind their peers?
If they do use it, is that naughty? Will they be caught cheating? But most importantly, and this one is happening a great deal, if they don't use it, will they still be accused of using it? And there are now young people who might use the tools, and then they work really hard to make it look like the output wasn't done by tool. They introduce spelling errors and grammatical errors just to confuse the world.
But we know that around the world, young people are being removed from their courses, having been accused of using these tools. So the picture is incredibly complex. But, again, I will leave it at, by using these tools, by promoting these tools, by trying to integrate these tools in education today, we are experimenting on children, and we're doing so outside of any ethical controls whatsoever. That's your choice. My choice is that’s not a good thing.
Speaker 6
Yeah. Organize their classes to inspire their children, their students, etc. So I believe that the first thing we need to do is empower these teachers. We have done studies in many schools in Europe, around Europe, in the whole world, right, everywhere. So, teachers are afraid.
Teachers are that they don't have control anymore. Right? So I think the first step we need to do is to empower teachers, to tell them that they have the most important tools already there. It's their pedagogical tools. If you organize your class, your projects, etc., around certain pedagogies.
I mentioned previously constructionism. Constructionism means that we don't have consumers of AI. We have people that interact with AI tools. Still, they interact, but they interact as designers, as doers. So this is a completely different story.
Through projects, for example, we have the project-based learning for many, I mean, Jews. It was like the previous one in the beginning of the previous century. It was like learning by doing. If we just empower teachers to use their tools, I think already we will have done a great step into the use of AI in schools. I know that I mean, students, professor Holmes mentioned, they will use it anyway, probably not inside the school, but they will use it afterwards.
So I think our role is to make students and to inspire students to understand the value of using AI in a way that they will benefit. Right? And probably to make them critical and to make them understand that they have the option of not using AI and how not using AI, they might benefit as well. So my first point is empower teachers and tell them that, I mean, even now with their current skills, they have the power to organize their classes in a very powerful way. Then, students.
Students is very challenging because most of the times, they are much faster than their teachers and their parents in adopting new tools. So there, again, I mean, it was mentioned beforehand about critical thinking or metacognition, etc. So there are ways, again, to inspire students to give emphasis to this. And for me, the most important part is we need, as you are policymakers, and we need mechanisms of evaluating the tools that will enter our schools. We cannot use whatever, I mean, our colleague previously mentioned that general purpose AI and AI that is specially designed for schools.
Right? So this is, I think, something that we should put emphasis on to have specific mechanisms on evaluating the tools. And, of course, we all have in mind that behind most of these tools, there are specific business models. And in fact, we are taking the brains of our future societies and putting them in these tools. Right?
So we need to be aware of this. Lastly, again, I will repeat the last part of my talk. I think that the collaboration between stakeholders is also crucial. Governments with industry, with researchers, with parents, with educators, with students, etc., should collaborate in a transparent way. Only in this way, we increase trust among stakeholders, and we find the best solutions for AI in education.
Thank you.
Thank.
Speaker 19
thank you very much. And remotely now, we'd like to hear from Professor Junta. You go ahead.
Speaker 7
Yes. Thank you very much. I will start with a question regarding the evidence. Do we have evidence that the frequent use of AI can hurt cognitive capacity thinking? Right?
Well, we will have evidence in a few decades. It's too early to have evidence because we just started, students started just started using consistently AI. However, we do have very strong suggestions, very strong insights. For example, neuroimaging studies show that the brain activation, when you study with access to an AI tool, the brain activation patterns are significantly different than the brain activation patterns that you have when you study without AI support. In the first case.
Speaker 19
Just wait a minute to see if you can be reconnected. I would suggest that we move on to the next point before we do have the next round of speech speakers. Perhaps we could turn to Mrs. Accente and ask you to start giving your presentation, and then hopefully get the reconnection back.
Speaker 23
Thank you very much, Chair. Are my slides up? Thank you, Branch. First of all, I want to thank the CALD committee for commissioning such an important piece of research and my colleagues for putting together such thorough proposition going deep into the subjects we've discussed today. My role is to summarize the key learnings of the workshop and provide some recommendations for the future.
There's an undeniable fact that AI is becoming a force that disrupts society. We've heard from the speaker on the panel and the excellent question posed by the parliamentarian that AI is disrupting not just education, but every single sector of society. Everyone is actually rather scared from my clients, the executive I work with, to parents, to children, perhaps children a little bit less because of various reasons we'll discuss in a second. It feels like we are flying a plane while we're building it, and that's an uncomfortable sentiment. Besides the use of AI in the classroom, Wayne was mentioning it has happened in the last ten years. And Irene has mentioned that children are one of the fastest growing user groups of such tools, the children are using AI outside the classroom as well.
There are some studies that show that 58% of them are actually using it for learning. 70% of them are actually using it for health advice. That was, to me, a little bit of a shocking recollection. But also companionship, 72 percent. So keep in mind that outside the classroom, the use of AI among children is rampant.
And what do you do? So if AI is here to stay and the genie is out of the bottle, we need to be able to support those who are designing, not just using AI, to have a better understanding of the needs of the users they serve. That's why this workshop and this research come at such an important time as an aid for everyone who's involved in AI education, from teachers to school managers to parents and carers themselves. I would want to put the parents on the map as well because they are an important stakeholder group often not part of the conversation. And especially when it comes to use of AI to shape the children future and the children character, they play an important role.
Part of the work of this workshop is not only to go deep into the three topics, pedagogical, cognitive, and ethical, but also being able to bring it together. The added value will be to create a unified framework that will support the stakeholders involved in this process. And, yes, we hear research after research. You know, the economy is calling AI a rewiring childhood. We've heard from Brooklyn institutions, Center for Universal Education, that the risk of AI outweighs the benefits.
However, there is a sense of optimism, and this optimism is coming from the fact that we are discussing how the pedagogical science is changed by the landing of AI. We've heard from Vicky that it's really important to ground the use of AI in educational goals. There is a temptation to buy the hammer and assume that we'll find the nails. This techno solution is it has been the approach the tech industry has been operated by. Just buy our product.
You'll find the solution. You'll buy, you'll find the problem later. And we have to be able to resist that. Really important is to be able to leverage these existing frameworks that have been developed by academia and industry. One, being human-centric AI.
Vicky proposes a student centric approach, teacher centric approach to allow us to center the design and the use of AI on the needs of the children, not just on the commercial interest of the AI companies. Also, Vicky is suggesting a greater focus on those capabilities, human capabilities that will not be touched for the foreseeable future by the intelligence machines, being the social emotional development, critical thinking, reasoning, meaningful social human interactions. Those will be the capabilities that machine it's going to be a no fly zone for the machine for the foreseeable future. At least, that's what I hope. The second pillar described, so eloquently by Irene, is around the cognitive capabilities.
And, we've learned that there is a learning paradox. It's the fact that AI, and I've seen this too many times, this AI paradox, on one hand, boosts some cognitive capabilities but erodes others. And with that, it comes four main categories of risk. Of course, those risks are not new. Right?
We've had digital platforms for decades now. Right? So, if we observe how we use certain apps in our phone, we'll soon acknowledge that we already are a bit overreliant and dependent on apps like Google Maps for navigation, and so many other tools that made us less adventurous and ready to use our own abilities to navigate in time. Of course, there are risks about skill impairment. It's enough to look at the abilities of teens to form a forge social relationship, not to mention erosion of autonomy or disruption of memory.
Again, those are not new risks. Those are that we've seen those risk emerging as an overuse and dependency of digital tools. We haven't got to the terms with social media, massive harms. And we know from the original design principles in Silicon Valley in the early 2000, there was a piece of design called persuasive design that was aiming to keep the user locked to the screen. The third pillar is, of course, the ethical dimension, and we've heard from Wayne, both in his presentation and answer to the question that there is a lot that can go wrong.
And we need to be able to anticipate it. We need to be able to address it. We need to be able to acknowledge that we have enough knowledge and a framework to be able to do that. There's a ton of quality work that has happened in conceptualizing the ethical dilemmas in relation to AI, and Wayne and his colleagues have worked to adapt that to domain specific as there are gaps and that more work needs to be done to make sure that this adaptation to education specific needs is done in a professional manner. But when thinking about AI in education and ethical dilemmas, we should not ignore the ability of AI to automate inequality. We've seen it in so many different cases.
If you go and search for AI incidents, we've seen how AI ended up automating discrimination of all sorts and kinds. So we need to pay close attention, intentional close attention to the digital divide. And I really like what Wayne was saying about literacy. It's not just about teaching people to use AI, especially for children. It's helping them navigate a world where machine intelligence will be all around them, and in that process, preserving humanity.
How do we bring it all together? Because each of those pillars should be considered, together and not separately. Pedagogy can help us define a purpose of what education is trying to achieve in a world with autonomous agents where we can benefit of personalizing information, providing incredible access to knowledge, to people from the global South, while at the same time paying attention to what AI, what's the AI impact on learning mechanism and also, of course, being able not just to identify the ethical dilemmas, but being able to manage it. Responsible AI in education means not just one of those pillars, because one without the other can create a series of other problems, but bringing it all together in a way that is not only practical but reduces complexity. Wayne was telling us, you know, what such a complex issue of AI in education is and its ethical dilemma.
So one of the challenges is how do we reduce complexity, build confidence, as Vicky said, and to be able to go ahead without losing faith that we are on the right path. What are the key conclusions that I want to leave you with is, first and foremost, go before tools. I talked about the tendency of the AI industry to say, here is a hammer. Buy it. You'll find the nails later.
We should resist that, especially in policy, and call out, and be able to understand that the reason we have such a strong push from Silicon Valley is because of their commercial interest. And, Vig, you were talking about the business models of AI. I think whenever we draft policy, we should keep an eye on how AI is being developed and monetized, not just developed and monetized. The second very important summary is being able to center the development and use of a tool on the child development pathway. In The UK, we have developed with help from Bib and Kindredan an age-appropriate design, regulatory framework that's being enforced by ICO, our largest regulator, and has inspired a similar framework in California.
Those are, this is a very good starting point to be acknowledging that children go to different stages in their growing up, and therefore, digital tool need to pay attention to those stages incorporated. In 2018, when I was in Silicon Valley, at the workshop on AI for children organized by the World Economic Forum and UNICEF, I was a bit shocked to learn from both Facebook and YouTube. They were saying publicly, look. We did not design those tools for children. We designed Facebook for kids, YouTube for kids.
But in the end, the children ended up using the big platform. So we don't actually have any consideration for children in the design of the platform. And it's enough to look how OpenAI is failing our children as well to understand the Silicon Valley is to catch up with the fact that children is not only the largest group, but is the future. And lastly, really important, and this is where I focus my area of work focuses mostly is being able to govern it. Now that we know there's a new tool that has a risk, we need to be able to be on top of it.
And to be able to be on top of it, recognizing that this is a tool, this is not an esoteric item, neither a demigod, but sometimes we kind of, you know, sense that the narrative is coming. It's a tool that has to be aligned with human needs, and it has to be governed accordingly. Where do we go from here? So, yes, there's a lot of complexity, a lot to be done. We are flying the plane while we're building it, but there's a way forward.
Part of the industry has been working with public and private sector in The UK, have managed to find a way forward without being terrified that there is too much to do or there is this sense of inevitability. First of all, is having been able to develop an integrated framework. Talked about this. I'm not gonna go too much into details, but say, being able to bring the three elements together and provide practical steps that could be applied right here, right now by the schools and teachers and parents around would be a great support. The second one linked with the fact that we need to be having more evidence on how digital affects the development of child at different stages is being able to build that corpus of data.
There isn't enough evidence. Biba Kidron with her Five Rides Foundation has done some work, but more work needs to be done. And lastly, but not least, last but not least, is the importance of leveraging best practices. We already do a good job of governing and controlling AI in other parts of society and industry. We should be able to leverage that and learn how to best procure it, ask the right question when those tools are being sold to schools, being able to understand how to best safeguard it, but not forget accountability.
When something goes wrong, and eventually it will go wrong, we need to be able to understand who's responsible. I want again to commend my three colleagues for their excellent work and for bringing those issues to you, and leave you with the last thought: despite the narrative that's coming so strongly from Silicon Valley and pollutes to a certain extent every single newspaper and news outlet out there, AI is a tool, a tool that needs to augment human capability, not replace it, and support the flourishing of the children. Thank you very much. Thank.
Speaker 19
you. Thank you very much. Barzo Pinesfuh.
Yes. Thank you. I'd like to thank all of you. There is now an opportunity for the group coordinators to take the floor. Again, I don't know.
Well, I'll give the floor to anyone who wishes to take the floor. Hannes Heide. Again, no? Milen Princi? Aletti?
Mr. Malczyk, go ahead.
Speaker 10
Well, it's such an intriguing discussion we have. I must not miss this opportunity. I see two very important issues, and I don't know whom to ask these questions to. But these are pressing questions as follows. How can we make sure that member states will not move at different pace and along different lines when the legal framework is being prepared?
Education is a national prerogative. So are we sure that uniform standards and criteria will be applied in all cases, at all times? I hope that this will be an important topic during the Cypriot presidency. I'm fully aware of the pressing geopolitical issues of the day, but we must not forget what we heard. We are flying an aircraft that we are in the process of building.
So thank you for all the information we received, and I very much hope that AI will remain just a tool, and not a substitute for human knowledge. Thank you.
Speaker 24
Yes. Thank you, dear colleagues. This was extremely interesting, very passionate. Thank you for your presentations. The challenges are huge, and it's extremely important, especially in terms of cognitive risks.
Mrs. Junta said quite rightly that the challenge is not to ban AI, but to make sure that we create the right kind of tools. Ms. Aksenti, you did say something about this, but perhaps you could further develop this and say something about the kind of tool we could create. In order to face up to these challenges, perhaps we could increase awareness about the fact that very much happens before the age of six, especially when it comes to language learning and logical reasoning, the kind of things that normally happen before the child gets to the primary school.
The use of artificial intelligence makes me think that when I was studying, for example, the teachers or lecturers allowed or not written documents. So perhaps others are you when it came to doing exams? Perhaps we ought to think of a way to allow students to use their neurons so not to have computers in the classroom or the exam room, I. E, make sure that they learn for themselves rather than using artificial intelligence. You've talked about policy and politicians.
I'd just like to remind you that it's the politicians who introduced screens into classrooms. And, this also was to allay a concern of parents, because they wanted to make their children's school bags lighter, so they didn't have to carry so many textbooks with them. Obviously, you can give examples of your own, but it does make me think of the calculator. At a given moment, there was a generation of students who were allowed to use a calculator in the classroom and then all of a sudden weren't able to do mental arithmetics. And obviously, those who never learned how to do mental arithmetic can't calculate using their own brains anymore.
Speaker 25
And this is what happens when you have these kinds of tools, because what it does do, it increases, exacerbates the digital divide, the digital gap, especially for the less well off parts of society. Thank you very much. We will now have an opportunity to hear from our experts again. I think Professor Junta is able to join us again from Japan. So perhaps, out of respect for her, we'll start with you again, and we hope that your connection does not break down.
Go ahead, Professor Junta.
Speaker 7
Yes. Let's hope so. So thank you so much. I will take advantage of the very last statement that was made. And I totally agree with this statement in the sense that if I want my students because I'm a teacher as well.
Right? If I want my students to use their neurons, then I design the task in a way that AI tools or computers will either not be useful for them or they will not have access to them. For example, we were using computers for programming education. Now I'm enforcing use of pen and paper. Right?
So I think that a big part of the solution is in the pedagogy that we are using, in the schools rather than in regulation, however regulation should exist. And I will go back to the evidence that we have regarding the use of AI and how it might hurt thinking, and we don't have hard evidence yet. We have good, strong insights. We know that it will be different. The cognitive architecture by consistent cues of AI, most likely will be different.
We don't know if it's going to be better or worse. Right? They might end up having superpowers in the next 100 years, our children, but we know that it is changing. And, therefore, we should be cautious, and I think this is about it. Thank you so much.
Speaker 19
Yes. Thank you very much. And now the last round of replies for our experts. And perhaps, one point you could address, I have the impression that the debate has been somewhat pessimistic, quite pessimistic. We've heard a lot about threats, pressure, and further challenges ahead of us.
And we as politicians are expecting from you some information about what we should be doing, what direction we should be heading in. And the kind of image that you have provided us with is actually quite pessimistic, as I've said. We know that artificial intelligence does provide possibilities for others brave. It weakens knowledge that has already been acquired. And it leads us to a kind of metaphysical laziness.
And as to whether it's harmful or not, well, we'll only know that in a few years' time. And the problems are increasing, challenges for parents and schools. And I think what is extremely important is that any kind of changes that are going to be proposed to pedagogical content have a specific structure. Do we have to change the objectives, the targets of education? If so, we have to have an overall idea of where we're heading.
So perhaps you could try to address those issues. So you have three minutes to reply. But I would like you, first of all, to thank you very much for your replies, but I would be very grateful to you as experts that you give us as politicians from the European Parliament some good advice, some good teaching, so that we're not trying to catch up, but that we're ahead of the trend as it were. Thank you.
Speaker 3
And thank you for the questions and the ideas. I wanna start with the notion of the calculator. I mean, it's one of the stories we hear a lot that the calculator came and it upended mental arithmetic. That there's some truth in that. But what we're seeing at the moment is the differences with these tools is that with a calculator, you press the right buttons, you get the right answer.
With these tools, it's a chance. Sometimes you do, sometimes you don't. For young people, developing their ideas, and that's very different. But I would take it even further than that. You know, AI is actually different to all the technologies we've seen so far, and it's different because it appears human like.
Speaker 4
It's sold to us as being human-like. It has this, we perceive it as being human-like, and that changes things. It means that it's far more powerful. Want to address the pessimism. I'm really sorry, but I am currently very pessimistic myself.
But what do we do about that? Well, the two things we should be doing are regulation. The metaphor I think of is in health. We don't go to our local shop and just buy some pharmaceuticals that was sold there, the local market. We expect our pharmaceuticals to be properly regulated, to be properly tested over time, robustly.
You know, there's lots of regulations about that. We also expect them to be given to us by an expert, a doctor. But all of that goes out the window when we're talking about these new tools that are so powerful, theoretically. Anything goes. We don't have any regulation.
But we also need the other point, the AI literacy, again. But this is focusing principally on the human dimension of AI, what AI is doing, how it works, how it's impacting on humans, how it's impacting on the environment, how it's impacting on human relationships. And one of the things there that's very important is that we should be empowering our teachers and our young people to be able to reject AI, not just assume AI is good and it's gonna be used everywhere, but to give them those capabilities and say, well, hang on a second. No. No.
Speaker 3
No. No. This is not appropriate for what I want to do at the moment. And the final point I'll make is with regards to funding. If I would go to a funder such as the European Union and say, hey.
I've got this really good idea for developing an AI-embedded tool, and it's gonna transform education. Then I often will have open doors. People want to listen to what I say. But if I go in and say, well, actually, there are problems here, and I need some funding so we can explore these problems to work out. How do we address these problems? Well, the doors close very quickly.
Speaker 4
So we need our funding mechanisms to be more open, so that we can properly understand this stuff. I've given you my opinion based on the evidence I have seen, but as everyone said, that's not enough. We need to understand this better than we do now. So, yeah, we need funding for that. Thank you.
Speaker 6
Yeah. I think I will tackle a similar point as professor Holmes and build a little bit and go one step further, Gundigal calculator indeed. But there is one more point there which is very unique to generative AI systems, and is that we come into a dialectic relationship. We discuss with these tools, and this is extremely powerful. We know that there are assumptions, biases, etc.
So through this dialectic relationship, children develop, they formulate their thoughts. Right? It's not just press a button for the right number. Right? So it's, like, even more and very worrying.
Second, about the pessimistic aspects. Here, again, I think we all know that AI has provided the tools for medicine, for example, to have really striking results for sciences, etcetera. So we have seen examples in science where things accelerate, and this is really amazing for human societies. But there you have experts using AI. I think it was mentioned before, we need to focus on building these cognitive capabilities in children in order to use AI in the best possible way, and we don't say that AI is not useful.
Right? And the last one is about funding again. And here, I want to mention another point. We know, not in Europe, probably in Europe, but worldwide, a lot of funding in education comes from industry. And there, I believe we have a conflict of interest.
Right? So we know OpenAI or Facebook or Meta, etcetera, they provide money to researchers, right, and to do research. We would love to see, as researchers, funding from an independent institution like a governmental institution like yours, the European Union, like the European Commission, etcetera, which does not connect to industry. And there, as it was said previously, to be open for any kind of result. And this is something that we really need because we need evidence about the positive or negative impact on children.
Thank you.
Speaker 23
Thank you very much for your question, Mr. Farooqi. I think it's a very good challenge, is where do we go from here. And I will remain an optimist because I've been working in AI now for more than a decade, and I've seen some major progress in how everyone in the ecosystem has come together. So to your question, I think the first thing I would recommend is to create some practical guidance.
I think a practical guidance that's easy to implement right here, right now, in schools will take everyone a long way. There's a ton of work that already happened. There are international organizations, think tanks that have produced such guidance. It's just a matter of putting it together and providing it to the school. It doesn't have to be something sophisticated.
It needs to be something that's workable. The second one is about experimentation. The U. K. Regulators have come together to say, look, we don't want to overdo regulation so that we don't stifle innovation.
Let's create sandboxes. So the concept of sandboxes, I think, is extremely important, being able to bring the AI companies, both general purpose AI or specific AI in education, and experiment alongside the students. And that leads me to actually the point around why I'm optimistic. Since 2018, I've been working with UNICEF on a program called AI for Children. And part of this program, we've created a policy framework, an ethical AI policy framework, but it looks at the needs of the children above and beyond education.
So consider the child as from a holistic perspective. But also we've interacted with children from around the world. We organize workshops in the global South, in Europe, in Asia, in Africa, and we've seen the potential involving the children has to give us a boost. The children understand AI much better than us, the adults. Study after study shows how children understand AI better than their parents.
Seventy percent of them were able to say, I know AI hallucinates. Therefore, I am careful how I use it. I think that this is amazing that a child of, you know, a young age is able to say that something that the experts are articulating barely. So the reason for optimism is that the next generation has a very different understanding of the digital world. They grew up in it.
They are playing or spending most of their time there. So I will encourage the European Parliament to create the right structure to engage the children, not just in their capacity as pupils, but as children. You'll be amazed how much you can uncover. To your question, missus Sorrell, about the fact that politicians are the ones that are pushing those tools, I think the recommendation I've just provided might alleviate a little bit. I think the reason they do it is because of the lack of knowledge.
What does this tool do, what are the commercial interests of the AI companies? Again, we go back to literacy. It was the case of a country in Europe, not giving names, that has just signed up a partnership with OpenAI, the same company that had to wait for a couple of cases of children analyzing themselves to actually acknowledge how much children use AI. So I think, again, being able to bring those companies and bring them on a journey that will see them acknowledging the children as users and take the existing frameworks that look at the children needs throughout the development life cycle and build tools with children in mind will get us a long way. At the moment, what we have to address is the Silicon Valley, not even today, are acknowledging children.
They're still waiting for disaster to happen, and then they are responding to the disaster rather than proactively do it. And I think European Parliament can play such an important role as have done it in so many other instances, to lead the way and bring those companies alongside and provide the right context so that children are not an afterthought. But first on the list when user design is being considered. Thank you very much.
Speaker 29
Ginco Barzo. So I will speak in English. The commission thanks the court committee and the members of the European Parliament as well as all the speakers for this very interesting and engaging discussion today on the topic of AI in the classrooms. The three dimensions scoping today's discussion, ethical, pedagogical, and cognitive, are crucial for the successful and effective response of European school systems. However, they are not sufficient.
Speaker 28
I would like to add also the question of sovereign technology in AI for education, because if we are not in control of access to data, compute, and technology, then it is unlikely that we're going to be able to address the three other very important questions. Moreover, the reality is that AI is already a key part of the digital practices of students and teachers. Miss Akcenda just mentioned children are among the most innovative users, as a matter of fact. Young people are among the most innovative entrepreneurs in the digital space. So we shall listen and also see what they do, and trust that they may come up actually with creative solutions.
Then European digital, sorry. Digital Education Eurobarometer of December.
Speaker 28
The Digital Education Eurobarometer of December, which is a statistically representative study of 27 member states, confirms that 80% and more of EU citizens believe all teachers should be equipped with the skills to use and understand AI. So the question of adoption is not on the table. The question of evidence is a correct one, but as I believe Miss Karizi said, on some of those topics, evidence will come a little bit later, just like with adoption of major technologies we have seen over the centuries, such as cars or industrial transformations. Now the developments that are described here are calling for a strong and coherent support system. We see that the discussion is very far from being black or white, so there's a lot of nuance that has to be introduced here through evidence and research because it is evident that AI tools do open a whole range of opportunities for personalized and innovative learning as well as for disadvantaged groups, for groups for people with disabilities as well.
Voice is a very good example of that. And then we see, of course, a major issue with transparency, biases, and stereotypes, not to speak about disinformation and cyberbullying. But what is important is that we do not throw all of this in the same basket and say AI is the major problem here. Due to the AI tools which design, which resemble human behavior, we also have to pay very close attention to AI's socioeconomic, socioemotional, and cognitive impact. Full agreement on this, and we fully agree with the need for further research on this.
I am reminded of an example somebody told me about cognitive offloading of the fact that I, in the 80s, knew hundreds of telephone numbers. They, I don't know any. Is that good or is that bad? I leave it up to maybe all those of you who remember those times too. I think it's not that bad that I don't remember them today.
But the adaptation of the education system to the AIH is a matter of a strategic political priority. So I think this is evident. So it's a question of social fairness, democracy, and digital sovereignty too. And we stand to our commitment to promote the use and integration of AI in a way which is human centered, evidence based, and purposeful. I would like to finish with three references to some of the speakers, or to some of the examples that have been made here today.
First of all, the general guardrails, as has been said very clearly, are set by the AI Act. And in the last years, we have put a lot of targeted effort into translating what that means. Now it has been mentioned that it has not been sufficient. We agree. For that purpose, first, we have published the ethical guidelines on education for teachers.
Let me say that this was done before Chat GPT came. So, it is not that the commission has followed a trend, but rather has actually come out before that trend appeared. And this is because the commission has invested many years of research in the role of AI more generally in education. Incidentally, this document is among the most downloaded documents of last year on the European Commission website. I don't think it's because it's made in a TikTok.
It's just very interesting for teachers. So that's the first point, and it is very strongly rooted in ethical principles. Second point, it has been mentioned, and I thank all the speakers for the excellent points on AI literacy. The European Commission, together with the OECD, has published, here I have to correct, a draft of an AI literacy framework last year, which has been now in consultation for more than ten months. Very big, massive, inclusive consultation of multiple hundreds of thousands of stakeholders, and we will propose this framework in June, in a public event.
Let me mention that this AI literacy framework has four dimensions, and two of them clearly address the need to go beyond the use of AI. One, about creativity and dealing in a creative manner with digital content, with the digital world. Second, the shape dimension, which is about understanding the impact of AI and what it does mean in technical terms. Because unless we understand the technical part, it is, again, unlikely that we are going to be able to design or shape its principles. And then the last point that I would like to conclude with, and with the reference to Miss Hoot Hunt, Professor Hunt, I think, who has said that regulation is not enough and that it is teaching pedagogy and curricula that actually matter.
And I think, this is the point of evolving policy, which she mentioned. We very much agree with this. This is why we, as a commission, support the member states to address these types of questions as a matter of priority, and we thank the European Parliament for raising this attention on a very clear need to act in a purposeful, balanced, and yet decisive manner. Thank you.
Speaker 19
We also have a representative of the Cyprus presidency. I'd like to hand the floor now to Miss Kika Katami.
Thank.
Speaker 32
thank you very much. I would like to warmly thank you for inviting us today as the presidency of the Republic of Cyprus in this extremely important workshop for this very important matter. It is a great joy for me to be here today. And I would like to point out that apart from our distinguished speakers that enlightened us in terms of many of the important aspects of this matter, there are also three EU institutions represented here today: the Parliament, the Council, and the European Commission that are here to put their minds together on this very important matter using AI and education. And let me move forward in English, if you may.
Intelligence is increasingly becoming part of our education systems and, of course, of our broader societies. Whether we approach this development with enthusiasm or with caution, one thing is clear: AI is here, and it's going to stay. The real question for policymakers and educators is therefore not whether AI will shape education, but how we guide this transformation in a way that strengthens learning while safeguarding the human values at the heart of education. The opportunities are significant. AI has the potential to support new forms of learning, to enrich teaching practices, and to help education systems respond to the evolving needs of learners in a rapidly changing world.
At the same time, the integration of AI in education raises important questions that we cannot ignore. Issues related to ethics, transparency, trust, data protection, and the broader implications of how knowledge is produced and understood must be carefully considered, as already referred to by the experts. Ensuring that these technologies are used responsibly and in ways that respect fundamental values is, therefore, essential. For the Cyprus presidency of the Council of the European Union, these questions have been at the center of our work in the field of education. In this context, we are currently advancing council conclusions on teachers in the era of artificial intelligence.
Our starting point is simple but fundamental. Technology alone does not transform education. Teachers do. As AI becomes increasingly present in education and training systems, teachers must be equipped with the knowledge, skills, and confidence to engage with these technologies critically and responsibly. Empowering teachers to understand AI, to assess its pedagogical value, and to guide students in navigating an AI enabled work is therefore key.
And I want to admit that it feels very good for the presidency to understand that what we have heard before by the distinguished experts is being reflected in our political text. We emphasize teachers' agency. We underline the need to retain professional judgment, the choice of pedagogy, to be careful with the digital divide, and to advance AI literacy. We also speak about the need for professional development of the teachers in the use of AI. We try to develop a common vision among the member states for the council to highlight and focus on the role of teacher in the era of AI, to maintain the human dimension of teaching and learning, and to make teachers enablers on this endeavor to benefit from innovation while ensuring and promoting quality, equity, inclusion, and success for all.
Educational decision should remain grounded in pedagogical purpose. Addressing these developments requires cooperation across Europe. Artificial intelligence education is not a challenge that any single actor can address alone. It calls for dialogue between policymakers, educators, researchers, and technology developers, and for continued exchange across the European community. Only through such collaboration can we ensure that technological innovation evolves in a way that supports educational goals and reflects our shared values.
Let me conclude with a simple conviction. Artificial intelligence is steadily becoming part of the educational landscape. Our responsibility is to ensure that it strengthens rather than replaces human agency, professional judgment, and the educational values that lie at the very heart of teaching and learning. Thank you very much.
Speaker 19
Thank you very much. If I could briefly summarize, I'd like to start off by thanking all the experts because I have written a lot of meeting. Some were predictable, some were boring. But here, this was extremely interesting, and we did cover a lot of highly sensitive issues. When you have this clash between optimistic and pessimistic views, that gives us an opportunity for success because there are those people who can see success, can't see any obstacles or problems, and there are those inversely who see problems and obstacles and not so optimistic.
So I think we have had an interesting coming together of opinions. And, I think, yes, we can make progress. But on the condition that everybody involved, the parents, politicians, teachers, obviously, are involved, can heed advice. I think when it comes to children, they seem to be able to dominate the risk. And I think we have to be very sensitive, but we also have to be extremely critical so that we can make sure that it's the opportunities that come to the fore rather than the threats.
Obviously, we know that there are challenges, problems ahead of us. We know that there are attempted suicides by young people, and this is a fact. It's a reality. So that is a threat, and it's up to us to be responsible. We have a responsibility towards the younger generation, and that responsibility is growing.
And I would hope that our politicians can have the same kind of imagination as our experts. E, an airplane that is flying and being built at the same time. Well, that's yes, but I'm very happy that we have met people who have this kind of creative imagination. Very many of the information that you have given us has been noted. And obviously, we are going to try to find the necessary sources of funding for research.
It's very important, obviously, what has been said by the representatives, both of the commission and the Cyprus presidency, either this is an issue that is going to be continued or we're going to develop it further. So thank you very much to the experts. You've looked at pedagogy, a critical view of AI. And so thank you very much, and I would very much hope that we will be able to host you again in the European Parliament. Have a very pleasant evening.
And then to the members of the committee, we still have one point left on our agenda. So thank you very much.
Cristo Petrov, ready? Okay. For.
Ladies and gentlemen, we only have 29 left.
We must start it.
Speaker 19
I would just like to take the opportunity of this slight chaos to thank our interpreters. It is true that some of the presentations were quite fast. They were very intense. But you, as always, coped. So I would like to thank you very much for your work.
And this brings me on to point 11 of the agenda, implementation of the council recommendation of 30th October 2020 on a bridge to jobs, reinforcing the youth guarantee and replacing the council recommendation of 04/22/2013 on establishing a youth guarantee. And we're going to be considering a draft opinion. The rapporteur has said that he's ready. Mister Petrov from Renu, go ahead.
Hristo Petrov
Thank you, chair. Dear colleagues, the reinforced yield guarantee remains one of the European Union's most important instruments for supporting young people in continuing their education as well as in transitioning into employment. Its reinforcement in 2020 was necessary and welcome. It broadened the age groups of the scheme so as to better reach vulnerable young people and place greater emphasis on the quality of offers. It has contributed to lower youth unemployment and stronger policy coordination across the member states.
However, important implementation gaps remain. Too many of the most vulnerable young people are still not reached by the youth guarantee. And in several cases, we see imbalances between the different types of offers, with heavy emphasis towards employment even when education or training may offer a more sustainable pathway to the labor market. And that's why the opinion focuses on strengthening the educational, preventive, and inclusive dimensions of the yacht guarantee. The EPRS study on the reinforced shows that tailored education and training can help young people stay in employment further down the line.
This is especially true for youth with fewer opportunities. We can strengthen this by improving career guidance in schools and by training teachers and educators to better help young people navigate choices for their future, whether that is VET, apprenticeships, further education, traineeships, or work. Prevention of early school leaving is another element that should be strengthened. It requires good cooperation across schools, social services, and employment services so that young people don't leave school due to socioeconomic reasons such as lack of transport or living in poverty. And this cooperation is also necessary to provide an immediate response in the event of dropouts.
This is also where integrated service models and one-stop shops have shown promising results, and we should encourage their use by more member states. Finally, the opinion underlines the contribution of culture, the arts, sports, and youth work. They are not simply tools for outreach. They are formative environments in their own right. They help young people develop soft skills, basic skills, creativity, responsibility, confidence, and civic and social engagement.
For those who do not or cannot engage with public services directly, these trusted and low-threshold spaces can provide essential intermediary steps back into education, training, and employment. Recognizing the learning that takes place in these settings is therefore crucial, and that's why validation of nonformal and informal learning matters, especially for vulnerable groups. To conclude, the reinforced yield guarantee must continue to evolve and fully embrace its educational pillar. Only then it will become a framework that is more preventive, more coherent, more inclusive, and better equipped to offer young people across the European Union the chance to build successful lives. Thank you.
Speaker 8
Thank you, chairman. The youth guarantee is a pillar of the European Social Fund, which the EPP fully supports and has done since 2013. It provides concrete opportunities to more than 24,000,000 youth and contributes to the reduction of the rates of unemployment. But there are still some critical aspects we cannot underestimate: school abandonment, persistence in problems in the labor market, the disparities from territories, in particular, rural and ultra peripheral areas. For the EPP, the success of this instrument can be measured according to three main conditions.
The key of qualified education, national responsibility accompanied by targeted European support, and capability for building synergies between education, families, and social infrastructure. With this in mind, we look favorably on the amendments contained in the report, which reinforce the specific efficiency of the instrument, preventing early school leaving in line with the European objective of reducing this 9% threshold. The attention paid to the most vulnerable groups, including young people in the rural areas, valuing the informal learning, which is essential to fill in the gaps in certain areas. At the same time, we feel it's necessary to avoid the text losing certain aspects of emphasis on youth unemployment as well as avoiding unmeasurable factors run the risk of undermining the nucleus, which remains access to a qualified opportunity within four months even through personalized services. While we recognize their value, we still need to make sure that there is certain accountability on the member states, because otherwise, the risk is that we would worsen the administrative burden without achieving any kind of results.
And this PRS assessment confirms that action is not homogenous, and support for the most vulnerable continues to be very difficult, and the four month target in many cases is not yet respected. This is why the EPP would stress one specific point. Instruments and tools are not enough. We need to understand what we've got already. So we would highlight three main aspects.
First of all, a stronger alignment between training and the labor market so that the young people's skills more effectively respond to the real needs of the companies, more decisiveness in preventing early school leaving, and support for the most vulnerable young people through instruments that are capable of intervening early and combating the risk of exclusion. Three, better territorial cohesion and a fairer access to European opportunities, with particular attention paid to rural and ultra-peripheral regions where young people remain further away from opportunities of mobility, training, and inclusion. Thank.
Speaker 9
You, chair. It is a pleasure and an honor for me to be able to speak on behalf of our shadow rapporteur, Nicolas Singoretti, who is unfortunately not able to be here. So, first, he thanks the rapporteur for their draft, who feels it is a good foundation for the work on the report. And I can just describe which direction he would like it to move in. We have to make sure that the youth guarantee can be effectively accessed by all young people who need it the most.
I thank the rapporteur for putting education at the heart of it and for young people, because it is difficult for young people in precarious circumstances to get qualifications. It shows clearly that an approach to quickly get people into jobs is not really going to be enough. It shouldn't be that the youth guarantee is oriented to short term needs of the labor market. We need a paradigm, vex. So it shouldn't be about trying to move people into jobs very quickly, but rather it should be for high quality jobs that last in the long term.
At the same time, we have to support people by helping them to support their health, because that will improve the situation for young people who are disadvantaged. They should be involved in the policy. This can only be managed by having better coordination at all levels, better data collection, better range of services, and support being offered. It should be unbureaucratic. They need support in the regions and in local communities, especially in rural areas.
Speaker 34
And I've noticed that it's very important to other people as well. Thank you.
Please.
Speaker 35
So, colleagues, the youth guarantee, well, it is a commitment to the young generation. We support these objectives completely. However, this guarantee has to be approached in a responsible and realistic way. So this report shows that young people, it's not only that they have to take part in this program, but they should also be prepared for the future. So they need high quality training and education, and they need to have good opportunities in the labor market.
So education plays a decisive role here. But that should not mean that only access is important, but quality is important, and development of people's skills and a sense of personal responsibility. Young people should take part actively in order to build a pathway to their own future. So the instruments that are being proposed need to be practical and efficient. We have to reduce bureaucracy, strengthen families in their role, and the role of the schools and the local authorities.
Of these, stakeholders should be strengthened because these are the stakeholders who stand closest to the young people. So with regard to the responsibility for implementation, that has to stay with the member states. The solutions should be adapted to meet the local and national state situation. Especially young people who live in regions where there are less opportunities should be helped. Equal opportunities don't just mean access, but they also mean fair conditions for every individual to develop themselves.
Whether the guarantee will be successful for young people. Well, it will be successful if it really does create long-term stability and opportunities for the next generation.
Speaker 10
Thank you, chair. One more time, I present our excuses that Madame Magoni cannot be here today. So on her behalf, I will read the note that she prepared on this opinion. So, mister president, honorable colleagues, the social fund must be a concrete instrument to support quality employment, vocational training, social inclusion, not an additional layer of bureaucracy or centralization. In this context, it is also important to strengthen the link with European investments in knowledge and innovation such as those foreseen under the Horizon Europe program for the period 2034.
Research and innovation play an essential role in developing skills, supporting educational systems, and fostering the growth of cultural and creative sectors. Employment policies work when they are rooted in local realities. It is regions, businesses, universities, and local communities that truly understand the needs of workers and young people. For this reason, the new partnership framework must ensure flexibility, administrative simplification, and genuine involvement of local actors, also enabling universities and
Speaker 10
Okay. Well, I'm not sure about my English. That's why I'm in a hurry to finish it. So I will try once again. So for this reason, the new partnership framework must ensure flexibility, administrative simplification, and genuine involvement of local actors, also enabling universities and cultural institutions to access the European opportunities more easily.
Investing in skills means strengthening Europe's competitiveness. It means supporting training, innovation, and young people's access to the labor market, with a focus on concrete opportunities. Only this way can Europe truly be close to its citizens. Thank you. And thank you.
Speaker 37
Thank you so much. Thank you, chair. And thank you to the rapporteur for this draft opinion and for the solid basis it provides. So from our perspective, we see this as a key opportunity to strengthen the link between education, culture, and employment policies, and we stand ready to work constructively together on this file. As Greens, we know that the impact of the youth guarantee remains uneven, with persistent gaps in access, outreach, and outcomes, especially for young people with fewer opportunities.
This shows that we need a stronger focus on education and training pathways, not only on short-term employment outcomes. And this also shows that we need to make sure that these opportunities are made clear and available and accessible to all young people, wherever they live or wherever is their background. And they're especially focusing on the people that are systematically discriminated and excluded from the job market and our educational systems. We also want to emphasize that education and training are not only about skills, but also about personal development, autonomy, and civic engagement, which are essential for sustainable integration into the labor market. We believe that young people need continuous accompaniment through career guidance, mentoring, and better cooperation between education, youth, and employment services, especially at key transition points and where the skills gap is the biggest.
We also stress that the importance of non-formal learning, youth work and volunteering help build confidence and transversal skills and must be recognized and supported. We look forward to working closely with the rapporteur and the colleagues to strengthen the education and youth dimension of this report, to make it more just, more accessible, and to deliver a more effective response for young people. Thank you so much.
Thank.
Speaker 16
You, chairman. Youth guarantee is without a doubt an important instrument to promote youth insertion in the labor market. It provides specific opportunities for first contact with the working world, but we shouldn't avoid mentioning the critical nature of this. The program doesn't work homogeneously throughout the European Union. There are certain areas where precarious and underpaid work continues, and this is unacceptable.
A tool designed to fight youth unemployment cannot be transformed into a mechanism which de facto increases it. A significant section of youth, about 60%, remain unemployed even after having checked in on the program. This is a serious threat to its sufficiency. So we feel it's necessary to reinforce the youth guarantee by increasing available resources, by introducing stricter criteria, would guarantee quality, accessibility, and equality of treatment in all member states. It's also important to promote sharing of good practices between those countries where the program has given positive results, to build a more coherent and effective approach throughout the union.
Thank you.
Speaker 19
Thank you very much. Unfortunately,
We don't have time for questions. And I'd now like to hand the floor to the commission for three minutes, Miss Chiaramonti. Please stick to the speaking time so that the colleague can reply to the comments, the rapporteur can reply to the comments made.
Speaker 38
Thank you, chair. Thank you, rapporteur, and good afternoon to everyone. So the Youth Guarantee has proven a fundamental instrument in supporting young people in getting work experience and developing the right skills for a changing world of work. The Use Guarantee has been a great success, as proven by the fact that since its creation in 2013, it has helped almost 57,000,000 young people through an offer of employment, continued education, apprenticeships, or traineeship. This is also in line with principle four of the European Pillar of Social Rights.
The principle which covers active support to employment more broadly stresses in particular for young people that they have the right to continue the education, apprenticeship, traineeship, or a job offer of good standing, and I really want to stress the good standing, within four months of becoming unemployed or leaving education. The commission believes that the use guarantee framework covers all the relevant aspects and is easily adaptable to different realities. And we heard in many of your interventions the importance of really being, you know, adaptable and applicable in each different reality. In the view of this, the commission believes that the use guarantee remains valid and up to date. The commission agrees that the implementation at member state level has been uneven, and this is due to different reasons, including different starting points.
However, the March 9 EBSCO meeting confirmed considerable recent progress in the implementation of the use guarantee throughout the EU, in particular, in countries where large gaps remain. The commission continues to support member states in their efforts of implementation. This includes, among other things, sharing of good practices, the annual data collection, and an EMCO review every two years. In addition, funding, in particular through the European Social Fund Plus and the RF, as well as monitoring of youth employment and needs through the semester, offer an important support to member states. We have read with attention the draft opinion to the draft implementation report proposed by the CALD committee, and several of the aspects highlighted in the CALD are already part of the reinforced use guarantee.
For instance, supporting young people with fewer opportunities, structured coordination between different actors involved in delivering the use guarantee, data sharing to enable timely, targeted and personalized support, and the focus on prevention, early intervention, career guidance, and soft skills provision. And just this is to name a few. On other aspects, we believe that existing or planned initiatives of the Commission are already addressing or will help address some of your concerns. For instance, existing council recommendations, such as the recommendation on micro-credentials, to name one, and other actions that are upcoming under the Union of Skills. In addition, we recognize the importance of volunteering as an enriching experience for young people.
And for this reason, in 2016, the commission launched the European Solidarity Corps. Nevertheless, volunteering usually does not offer a direct link to sustainable integration into the labor market, which is the goal, the main goal of the use guarantee. Moreover, the commission agrees that monitoring is essential to assess the efficiency of the schemes and to improve them, which is why we are regularly working with the member states to ensure quantitative and qualitative monitoring of the use of guarantee implementation. For the commission, it is essential that the use guarantee is fully implemented on the ground in order not to leave any young person behind. And for this reason, we are looking forward to working together with the parliament to improve the labor market prospects of all young people, and we are looking forward to the final opinion and the final report.
Thank you.
Speaker 39
Well, you have a couple of minutes if you need them, Missus Shubim Lam.
Hristo Petrov
Thank you, chair. Thank you to the shadows, also to the commission. It is very useful to have this first round of exchanges. And for me, two points are particularly important as we move forward. First, we need to broaden our understanding of what the Yacht Guarantee should encompass so that it better takes into account the soft skills and experiences that young people gain through culture, sports, youth work, volunteering, and community engagement.
We need to do more on promoting opportunities for the validation of non-formal and informal learning. Currently, procedures remain too complex, and the implementation of the 2012 council recommendation is quite uneven. We need to tailor them to the groups that need them most, such as vulnerable needs. Secondly, youth guarantee schemes should be expanded to include volunteering and youth work. They can provide meaningful experiential learning, especially for young people facing multiple barriers.
And here, member states do not need to start from scratch. The European Solidarity Corps already offers valuable experience in quality assurance, mentoring, learning outcomes, and transitions for young volunteers. And the European youth work agenda provides a shared framework for strengthening youth work across the EU. The timeline for this opinion is quite tight, but since we need to be ready by June, I believe we have enough time to produce a strong text. I think our work here can really help to shape the final outcome of this implementation report.
I look forward to working with you. Thank you.
Speaker 19
Yes. Thank you very much. Just a couple of formal points before you leave. I'd just like to remind you that the deadline for tabling amendments is 03/24/2026. That's twelve at midday.
And the amendment should be submitted in the digital system using the digital signature. The next meeting is going to be in April of this year, of course, and here in Brussels. And right at the end, I would like to thank all of you. Let's cross our fingers for the drafter of the opinion. And obviously, any amendments tabled have to be in line with the competencies of our committee, because sometimes we see that other committees table amendments that do go beyond the specific competencies of their committees.
So let's set a very good example, to make sure that our commitments are within our competencies. Having said that, cross our fingers for the rapporteur, and the amendment should be set for March 24 this year. Twenty-fourth. It's the 27th.
7. It's okay.
Yes. Because the gentleman said 24.
Ask anything about this video — I analyse its transcript to answer.