AI Giants Face Antitrust Heat Over Public Pact to Slow Model Progress
OpenAI said outside evaluators will test models during training and evaluation as it seeks earlier detection of safety risks.
- On Tuesday, OpenAI announced a framework for independent safety evaluations, allowing third-party groups to assess models during training rather than only before launch. The company is in talks with METR and Redwood Research.
- Industry leaders including OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei have increasingly called for stronger safety measures. Amodei urged firms to slow capability improvements, while Altman pledged to give independent evaluators employee-like access.
- OpenAI's approach includes 'grey box' access, allowing assessors to test whether safeguards withstand jailbreaks and prevent capability uplift in cyber and biological domains. The company prioritizes scientific rigour and security practices for outside investigators.
- Leaders are seeking government intervention to regulate AI and put more humans in the loop, though CEO Pankit Desai of Sequretek warned "they need to walk the talk." Amodei and Altman requested antitrust waivers for competitor coordination on safety.
- Intense business and competitive pressure complicate efforts to regulate AI development, as American mathematician Bailey Whitfield Diffie argues that controlling development speed is difficult. Some fear if the U.S. slows, China may advance instead.
31 Articles
31 Articles
OpenAI will let outside groups test its models during training
OpenAI says it will let third-party groups run technical safety assessments while models are being trained and evaluated rather than only before launch, and is in talks with groups including METR and Redwood Research. It has named no partner and set no access terms, ten days after Sam Altman promised evaluators employee-level access and the […] This story continues at The Next Web
OpenAI to Let Outside Groups Evaluate AI Models at Earlier Phase
OpenAI plans to let third-party groups vet its artificial intelligence models for safety risks in earlier phases of the development cycle, part of an ongoing effort to address heightened concerns about the technology’s potential harms.
Bizpac Review
thedailybs.comAI giants reportedly neared deal to safety test each other’s models
Two of the largest artificial intelligence labs held talks to safety-test each other’s models, according to a report. OpenAI and Anthropic negotiated a legal agreement to test […]
Coverage Details
Bias Distribution
- 62% of the sources lean Right
Factuality
To view factuality data please Upgrade to Premium
