AI firms discuss FINRA-style standards body amid safety concerns
OpenAI says it is working with Anthropic and Google on a proposed body to test powerful systems, but the companies have yet to agree on how it would operate.
OpenAI is working with Anthropic and Google on a proposed industry standards body for testing powerful artificial intelligence systems, as safety concerns increase pressure for stronger oversight.
The proposal originated with Google DeepMind chief Demis Hassabis and is modelled on the US Financial Industry Regulatory Authority, which oversees brokers and investment firms. Hassabis has suggested the body could test advanced AI systems before they are released publicly.
Hassabis said substantial funding would probably need to come largely from industry to secure technical expertise and computing resources for large-scale testing. The proposal’s funding model, membership, independence, testing methods and enforcement powers remain unclear.
France 24 reported that the companies have not reached consensus on how to proceed. The status of the proposed body, including whether it would be formally established, has not been settled.
The discussion comes amid renewed warnings about AI risks. Former Anthropic employee Jacob Coxon and former Google DeepMind researcher Bilal Chughtai have raised serious concerns about the industry, while Anthropic chief executive Dario Amodei has called for a slowdown in AI development.
Cohere chief executive Aiden Gomez has argued that AI needs guardrails but questioned who would write them, participate in the process and benefit from the rules. US President Donald Trump and Nvidia chief executive Jensen Huang have challenged predictions that AI could destroy humanity.
