Key facts
- OpenAI, Google DeepMind, and Anthropic are collaborating on AI safety.
- The companies have been in talks for several weeks regarding AI safety.
- OpenAI believes an antitrust waiver is not needed for safety coordination.
- The collaboration aims to address potential catastrophic risks associated with advanced AI technologies.
- OpenAI's global policy chief Chris Lehane confirmed the collaboration.
- OpenAI supports legislation requiring independent verification organizations for AI model safety.
OpenAI is collaborating with its rivals Anthropic and Alphabet's Google DeepMind on artificial intelligence safety, according to OpenAI's global policy chief Chris Lehane. The move comes amid rising global concerns about AI's existential risks, following a rare show of unity from leading AI labs over the weekend calling for a pause in AI development.
Lehane indicated that the discussions between the three companies have been ongoing for several weeks. He also stated that OpenAI does not believe an antitrust waiver is necessary for them to coordinate on safety measures, emphasizing that it is "better to try to work together to prioritize safety." Lehane was in Washington to meet with lawmakers and expressed OpenAI's support for any bipartisan legislation aimed at mitigating catastrophic AI risks.
The collaboration focuses on developing frameworks and protocols to ensure AI systems are designed and implemented with safety as a priority, aiming to address potential risks without stifling competition. Anthropic CEO Dario Amodei's recent essay called for industry collaboration to slow the pace of frontier AI development and avoid catastrophic risks, a sentiment echoed by OpenAI's Sam Altman, Google's Demis Hassabis, and xAI's Elon Musk. Altman stated OpenAI would join Anthropic in embedding third-party evaluators to monitor safety.
However, concerns have been raised about potential antitrust violations if the coordination is seen to suppress competition. While Amodei proposed a narrow government waiver for safety coordination, Lehane stated the firms do not need one. The Information reported that the three companies are also working to create an AI industry standards body. Meanwhile, President Trump has dismissed AI safety concerns as a hoax, and his advisor David Sacks echoed this sentiment.