OpenAI and Anthropic Slowdown Calls Trigger Antitrust Fears
As OpenAI and Anthropic call for a coordinated pause to address safety risks, legal experts warn that a joint slowdown could trigger severe federal antitrust investigations.
Major artificial intelligence developers are navigating a complex legal minefield as they seek to coordinate safety-related development pauses. Following reports of AI agent swarms compromising websites and warnings from researchers, industry leaders have proposed a joint "slowdown." However, antitrust experts warn that collectively agreeing to restrict development could violate the Sherman Act, drawing scrutiny from the Federal Trade Commission and the Department of Justice.
Legal scholars point out that how companies frame these decisions is critical. Roger Alford, a law professor at Notre Dame and former DOJ official, warned that a mutual agreement not to improve products could be viewed as "quality fixing." He pointed to a European case where automakers were fined roughly one billion dollars for colluding on emissions technology limits. Conversely, former DOJ policy director David Lawrence suggested that agreements preventing catastrophic risks are protected under the "ancillary restraints doctrine" because they ultimately promote competition. Alternatively, the National Cooperative Research and Production Act of 1993 offers a path to establish standard-development organizations if companies notify regulators.
Not all industry leaders support a formal pause. Meta CEO Mark Zuckerberg declined to endorse a slowdown, arguing that market forces naturally incentivize companies to align their models safely to avoid losing customers. This debate unfolds as Anthropic and OpenAI, which compete closely on the Ramp AI Index, prepare for public markets. Both companies have valuations near or exceeding one trillion dollars and have filed confidential paperwork for initial public offerings. Anthropic plans to go public next month, while OpenAI CEO Sam Altman has delayed its IPO until 2027 due to safety concerns.
For AI practitioners and engineers, this legal gridlock means that standardized safety protocols may remain fragmented. Without clear regulatory exemptions, individual labs must independently balance rapid deployment against safety risks to avoid costly, multi-year federal conduct investigations. These probes can force companies to turn over millions of pages of documents and subject employee devices to litigation holds, stalling actual technical progress far more than a voluntary safety pause would.
This is our own summary of reporting by WIRED AI
