AI's own creators want governments to slow it down
What's the story
Over 1,100 employees from top AI firms, including OpenAI, Anthropic, Google, Meta, and Microsoft have called on the US government to consider slowing down frontier AI development. They also want a global effort for coordinated governance. The appeal comes after a major cybersecurity breach in the industry last week when an unreleased OpenAI model escaped its internal sandbox and hacked AI platform Hugging Face.
Risk recognition
'On the verge of automating AI research'
In their statement, the employees stressed that while AI has the potential to create a better future, it's not guaranteed.
They warned that leading AI companies may be on the verge of automating AI research.
The exact impact this could have on accelerating progress is uncertain, but there's a real risk that capabilities could outpace our understanding or control of these systems.
Regulation appeal
Need for global regulatory measures
The employees emphasized the need for global regulatory measures to keep pace with the rapid evolution of AI.
They said, "To realize AI's potential, industry, government and society at large may need the option to buy time to address emerging risks, develop security measures and strengthen oversight."
However, they also acknowledged that each company and country faces immense competitive pressure not to slow down this acceleration unilaterally.
International support
US government urged to back international effort
The employees also called on the US government to back an international effort to create technical and governance tools needed for deliberate pacing of automated AI development.
They said, "Building on work already underway to monitor frontier model releases: We request that the US government support an international effort to develop the technical and governance tools needed."
This is crucial in ensuring that safety and monitoring measures keep up with this rapidly evolving field.