OpenAI is seeking clarity from Congress on the legality of a coordinated slowdown in AI development among competitors due to antitrust concerns.
OpenAI has approached U.S. lawmakers to clarify whether a collective slowdown in AI model development could breach antitrust laws. This development comes amid growing concerns about the rapid advancement of AI technology and its implications.
Industry competitors including OpenAI, Anthropic, and Google are considering a coordinated approach to slowing down their development processes. This raises potential legal issues, as such collaboration could be seen as an agreement to limit production, violating the Sherman Act, according to Nicholas Felstead, deputy director of the Australian Competition and Consumer Commission.
"The legal uncertainty itself can be a powerful deterrent," Felstead stated, highlighting the complexities of antitrust regulations.
Notably, John Schulman, co-founder of OpenAI, has emphasized that while antitrust laws restrict certain agreements, they do not prevent the collaborative planning of development timelines among companies.
Proposed Legislation for AI Collaboration
On July 23, a bipartisan group of legislators introduced the Collaboration on Adversarial Threats and Security Risks Act. Supported by Senators Adam Schiff and Jim Banks, as well as Representatives Bob Latta and George Whitesides, this bill seeks to provide limited antitrust exemptions for companies working together to address AI-related risks.
If passed, the legislation would allow firms to share information about threats and coordinate responses, including delaying the rollout of high-risk AI models. However, any coordinated actions would require prior notification to the U.S. Department of Justice's antitrust division, ensuring transparency about the risks involved.
As of now, the proposed bills H.R. 9914 and S. 5105 are in the legal committees of Congress, with no further developments reported since their introduction.
OpenAI's Previous Cautionary Measures
OpenAI's chief scientist, Jakub Pachocki, previously indicated concerns about the industry's ability to manage AI alignment and monitoring effectively, suggesting that a slowdown in development may be prudent until safety standards are established.
"I expect and hope that voluntary slowdowns will become standard practice until common safety standards are established," he remarked.
OpenAI had previously paused the scaling of new AI systems for two weeks following incidents related to Astra's cyber capabilities. This cautionary step reflects the company's acknowledgment of the fragile state of AI monitoring.
Moreover, the U.S. Senate has initiated an investigation into OpenAI's practices, particularly in light of recent AI agent behaviors, while proposals to ban the development of artificial superintelligence continue to circulate.
