AI giants seek antitrust exemptions to coordinate development slowdown
Leaders of major artificial intelligence firms are proposing a collective pause in model development to strengthen safety safeguards, a move that requires exemptions from competition law and has drawn criticism from rivals and foreign governments.
Executives from five leading artificial intelligence companies have called for a coordinated slowdown in the development of advanced models, arguing that current safety safeguards are insufficient. The proposal was initiated by Dario Amodei, chief executive of Anthropic, who released an essay this month urging peers to slow their pace. The position was subsequently endorsed by Sam Altman of OpenAI, Elon Musk of xAI, Demis Hassabis of Google DeepMind, and Satya Nadella of Microsoft.
The initiative has created a distinct divide within the industry based on commercial interests. While the proponents sell AI models or the cloud infrastructure to run them, Jensen Huang of Nvidia, which supplies the essential chips, has opposed the move. Huang argues that market forces are sufficient to drive safe innovation without new legislation. Mark Zuckerberg of Meta, whose company distributes models for free, warned that the proposal could expose firms to increased legal risks.
A central legal hurdle for the plan is the requirement to operate outside existing antitrust laws. Currently, competition regulations prevent these firms from holding private meetings to coordinate development pace. Proponents argue that exemptions are necessary for technological safety, but critics, including White House AI lead David Sacks, contend that this arrangement could reduce competition and allow the companies to inspect one another’s work. Sacks has specifically questioned the independence of nonprofit evaluators funded by the companies themselves.
The proposal also carries significant geopolitical implications. Amodei’s plan includes denying China access to the most powerful AI chips and advanced semiconductor manufacturing equipment, while preventing smuggling and remote access to computing capacity. China’s foreign ministry described these warnings as a fearmongering tactic aimed at slowing its technological advancement, a view Amodei acknowledged by admitting China poses the toughest predicament for his plan.
Supporters have drawn parallels to the Basel III banking regulations introduced after the 2008 financial crisis. However, a key distinction exists: banking rules were imposed by external regulators, whereas the proposed AI framework involves competing private companies coordinating their own pace. OpenAI has already slowed development following a security incident, and Altman stated the company will not go public in 2026 to allow for further safety advances.
The debate highlights a broader governance gap, as much of the discussion has occurred without meaningful participation from other regions, including the Arab world. With PricewaterhouseCoopers estimating that $31.6 trillion will be invested in AI infrastructure worldwide by 2050, nations with abundant energy resources may seek greater leverage in determining the conditions under which AI companies operate.
