Skip to content

Topic

AI lab governance

How decision-making authority and safety oversight are structured within AI labs building frontier models, including leadership roles.

Current stories

investConfirmed19 publishers

Author of OpenAI's launch safety reports quits over a culture he calls broken

David Robinson quit OpenAI after leading safety reports for 12 model launches and wrote in The Atlantic that the company's culture is broken. He helped write the safety process he now faults, so OpenAI cannot easily wave the critique away.

Perspective Coverage

23 publishers
Builder
Builder 30%
Operator
Operator 43%
Investor
Investor 27%

Reality

Evidence74
Adoption
Not scored
Hype gap+15
Incentives60
Confidence70
productOne report1 publisher

OpenAI's louder rollout made its independent math advisers look like company appointees

OpenAI's louder announcement of a nine-member independent math advisory group caused widespread confusion over whether the company had appointed it. Its first job is helping release scores more results from an unreleased OpenAI model, a wave researchers already dread.

Publishers:theverge.com

Reality

Evidence55
Adoption
Not scored
Hype gap+20
Incentives60
Confidence50
productConfirmed3 publishers

OpenAI dissolves the team that judged whether a model was too dangerous to ship

The Financial Times reports the preparedness team was disbanded last month, its work split by area into existing teams. Deployment risk models that assumed a vendor-side gate now need rebuilding.

Perspective Coverage

3 publishers
Builder
Builder 22%
Operator
Operator 41%
Investor
Investor 37%

Reality

Evidence60
Adoption
Not scored
Hype gap+15
Incentives65
Confidence60
investOne report1 publisher

Anthropic plan would hand seven founders 50.1% of the vote on roughly 14% of the equity

Anthropic wants its seven co-founders, who own about 2% each, to hold 50.1% of shareholder votes after its IPO, The Information reported. IPO buyers would get economic exposure while the founders hold most shareholder votes and an independent trust picks most of the board.

Publishers:crypto.news

Reality

Evidence40
Adoption
Not scored
Hype gap+10
Incentives60
Confidence40
productOne report1 publisher

Every task OpenAI lists for its new mathematics panel is advice

The nine members get early access to OpenAI's research, work unpaid and are free to publish what they tell the company. The researchers who spoke to The Verge still cannot say what happens when they disagree with it.

Publishers:theverge.com

Reality

Evidence58
Adoption
Not scored
Hype gap+28
Incentives72
Confidence55
productOne report1 publisher

Every AI safety control on this week's record was written by a company that sells models

Trump says the only guardrail AI needs is a strong and smart president. The controls a product team can actually cite are Anthropic's own plug-pulling, Microsoft AI's new outline and Monday's market selloff.

Publishers:bbc.co.uk

Reality

Evidence60
Adoption22
Hype gap+40
Incentives82
Confidence55
productOne report1 publisher

Brin shapes Gemini's model sizes and release windows from a converted microkitchen

Business Insider, citing eight current and former employees, reports that Google's cofounder holds no executive title while making technical calls on Gemini from a desk the new DeepMind SVP shares. Google declined to comment.

Publishers:thenextweb.com

Reality

Evidence40
Adoption
Not scored
Hype gap+14
Incentives62
Confidence38