Business, Innovation, Science and Trade Committee — hearing of 13 October 2026: transcript & video · Polyfact

Free trial Learn more

United Kingdom

Videos

United Kingdom

Learn more Log in Free trial

Business, Innovation, Science and Trade Committee

13 October 2026

Speaker dots

Speakers in the chapter
Colour: political group
Grey: no group

Tap or click a chapter's dots to see names and groups.

Contents updated on 13 Oct 2026

Committee hearing with independent experts on AI adoption, safety and governance, focusing on risks from advanced systems and implications for policy.

Layout

Layout Balanced Video, transcript and chat in equal parts Video Larger player Reading Wider transcript

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

Official source

Offset

Transcript offset

0s

Text size

Export

Select segments to export (Word)

0 segments selected

First panel — Independent AI safety and standards experts

21:51

Resignations from OpenAI and concerns about corporate responsibility for AI safety

21:51

Question from Rt Hon Liam Byrne MP to Daniel Kokotajlo on reasons for resigning from OpenAI

21:51

Question from Rt Hon Liam Byrne MP to Steven Adler and Daniel Kokotajlo on leaving OpenAI and internal safety culture

23:45

Question from Rt Hon Liam Byrne MP to Geoffrey Irving and Aengus Lynch on existential risk from advanced AI

28:32

Questions from Dan Aldridge MP on limits of alignment progress and industry risk tolerance

32:39

Question from Dan Aldridge MP on role of middle powers in setting AI safety limits

40:05

Question to Daniel Kokotajlo on OpenAI firings and whether company culture has changed

41:12

Question on companies’ duty to report lower-profile AI security incidents

42:00

Risks to democracy and concentration of political power from AI

43:42

Question on how AI firms address risks to democracy and elections

43:42

Daniel Kokotajlo on AI-enabled dictatorship and concentration of power

45:29

Testing, evaluation and governance of advanced AI systems

47:14

Question from Martin Wrigley MP on the reliability of AI safety testing and evaluations

47:14

Question from Mr Joshua Reynolds MP on who decides when AI models are safe to deploy

53:02

Question from Sarah Edwards MP on corporate responsibility for harmful AI behaviour

55:43

Question on how AI companies report safety incidents to investors and regulators

01:01:31

Questions from Alex Sobel MP on self-regulation and impact of external regulation

01:03:30

Question on best frameworks for AI governance and role of middle powers

01:09:50

Rt Hon Liam Byrne MP — summary of first panel's evidence and closure

01:15:26

Second panel — Frontier AI companies on AI safety and governance

01:18:37

Safety assurances for current AI models and pacing the frontier

01:18:37

Rt Hon Liam Byrne MP and Tom Duff Gordon (OpenAI) on guarantees that current models cannot cause harm

01:18:37

Rt Hon Liam Byrne MP and Lord Booth-Smith (Anthropic) on safety guarantees and pacing the frontier

01:21:34

Rt Hon Liam Byrne MP and Owen Larter (Google DeepMind) on safety frameworks, regulation and trust

01:25:16

Rt Hon Liam Byrne MP and Rob Sherman (Meta) on Meta's assurance that its AI models are safe

01:30:36

Committee questioning all companies on real-world cyber incidents and undisclosed misbehaviour

01:32:19

Alex Sobel MP and Tom Duff Gordon (OpenAI) on cancelling the Astra 6.1 release and alignment regression

01:39:33

Question on OpenAI's delay in notifying Australia and lack of alerts to governments on agent misbehaviour

01:41:08

Andrew Western MP and frontier AI firms on what risks keep them up at night

01:43:22

Dan Aldridge MP and Rob Sherman on MI5 warning about AI-enabled bioweapons

01:49:58

Committee clarifies companies' definitions of catastrophic and CBRN risks

01:51:34

Alex Sobel MP and Lord Booth-Smith on Anthropic's AI-linked wet lab and biosecurity

01:53:03

Questioning on OpenAI's delayed notification to Australia and loss of control over AI agents

01:54:42

Mr Joshua Reynolds MP with Owen Larter, Tom Duff Gordon and Lord Booth-Smith on closed models, AISI access and employee-like evaluators

01:57:46

Rt Hon Liam Byrne MP with frontier AI firms on statutory regulation, standards and international conventions

02:02:51

Question from Sarah Edwards MP to Tom Duff Gordon, Owen Larter and Rob Sherman on hacking notifications, model guardrails, third‑party evaluators and harmful imagery training data

02:09:42

Question from Charlie Maynard MP to Lord Booth-Smith and Rob Sherman on dropping pause pledges and controlling open‑weight models

02:18:00

Question from Rt Hon Liam Byrne MP and Charlie Maynard MP to Rob Sherman on MI5 warnings, kill switches and liability for harms

02:22:58

Question from Charlie Maynard MP and Rt Hon Liam Byrne MP to Owen Larter, Tom Duff Gordon and Lord Booth-Smith on pause pledges, kill switches and third‑party evaluators

02:27:08

Question from Antony Higginbotham MP to frontier AI firms on mandatory incident reporting and public disclosure

02:35:24

Question from Alison Griffiths MP to frontier AI firms on parity with US access, model testing and second‑wave UK AISI evaluations

02:41:38

Question from Alex Sobel MP to frontier AI firms on earlier‑stage involvement of UK AISI in model development

02:48:08

Rt Hon Liam Byrne MP — conclusion of second panel and introduction of government panel

02:49:56

Third panel — Government approach to AI safety, regulation and incident reporting

02:52:59

Government announcement on AI safety powers and use of the Cyber Security Bill

02:52:59

Question from Rt Hon Liam Byrne MP to Cat Little and Alex Jones on AI safety announcement, incident reporting and Cyber Security Bill amendments

02:52:59

Question from Charlie Maynard MP to Henry De Zoete and Cat Little on powers to block dangerous AI models and biosecurity risks

03:00:53

Balancing AI opportunities with safety and deployment in public services

03:10:25

Question from Sarah Edwards MP to Cat Little and Alex Jones on capturing AI benefits in public services while managing risk

03:10:25

Lessons from the July AISI incident and security of testing environments

03:15:24

Question from Martin Vickers MP on lessons from the July incident and robustness of AISI’s testing environment

03:15:24

Live capture

Versions of this video

Replay Final version · 13 October 2026
Live capture Streamed live on 13 October 2026

Current

Transcripts may differ slightly from one version to another.

Business, Innovation, Science and Trade Committee

13 October 2026

Committee Tue, 13 Oct 2026 Actual start 13:45 House of Commons

21:51 - 22:26

Speaker 1

Order, order. Welcome to today's session of the Business, Innovation, Science and Trade Committee inquiry into AI safety and governance.

We are incredibly grateful to you, our witnesses, for joining us here in the House of Commons today and for joining us at what I know is a very early hour on the West Coast. So thank you very much on behalf of the UK Parliament for coming and giving evidence to us. Daniel, I'm going to start with you, if I may. Could you just tell us why did you resign from OpenAI?

22:26 - 23:35

Speaker 2

So OpenAI and Anthropic and some of these other companies are trying to build superintelligence, which means AI systems that are better than the best humans at everything while also being faster and cheaper.

If you really internalize that and think about the implications for the world if they succeed, or even if they get anywhere close to succeeding, it's truly—well, there's so many implications, we can get into that later. But for years, these companies have been saying, "Yes, this will be incredibly transformative, yes, it could go really badly, but it could also go really well; that's why we need to be the ones doing it because we're the good guys, we know how to make it, you know, go better than those other bad guys out there who might do it less responsibly than us." Basically, I left OpenAI because I came to believe that OpenAI in fact would not behave responsibly around the creation of this technology, and that OpenAI was rationalizing why they needed to be in the lead and why they needed to build this as fast as possible. So basically, I lost confidence in the leadership.

The rest of the transcript is for Pro customers

Unlock the full transcript, chat, and every analysis feature with Polyfact Pro.

House of Commons

Watch the full session, transcribed live

Unlock word-for-word transcripts, AI summaries and chat on every parliamentary session.

Live transcript, every speaker identified
AI summaries and chat on every session
Alerts the moment your keywords are said

14-day free trial · cancel anytime

Free trial Learn more

Already have an account? Log in

Ask anything about this video — I analyse its transcript to answer.

House of Commons

Watch the full session, transcribed live

Unlock word-for-word transcripts, AI summaries and chat on every parliamentary session.

Live transcript, every speaker identified
AI summaries and chat on every session
Alerts the moment your keywords are said

14-day free trial · cancel anytime

Free trial Learn more

Already have an account? Log in

Keyboard shortcuts

Playback

Play / pause: Space
Back / forward 15 s: ←→
Back / forward 10 s: JL
Full screen: F

Chapters and transcript

Previous / next chapter: PN
Search the transcript: /
Turn follow on / off: S
Copy the current sentence as a quote: C
Show this help: ?

Single-key shortcuts J, L, N, P, C, S, / and ?. Turn them off if you use voice control or assistive tech that types letters.

Inactive while a text field, a button or the chat has focus.