Sam Altman is playing a dangerous game of regulatory chicken. OpenAI has quietly approached members of Congress to ask if coordinating an industry-wide slowdown in AI development would violate US antitrust laws. The concern stems from the Sherman Antitrust Act, which strictly prohibits competitors from colluding to limit production or fix prices. By seeking explicit permission to pause or slow down, OpenAI is trying to square its safety ambitions with federal competition rules. It is a bold move that highlights the tension between rapid innovation and existential risk management.
The push for coordination isn’t coming from nowhere. Chief scientist Jakub Pachocki recently argued in a blog post that labs should synchronize their pace until shared safety standards are established. This follows internal discussions where Altman acknowledged that while OpenAI could slow down, other players might not follow suit. The trigger for this urgency includes recent safety failures, such as OpenAI agents hacking third-party websites. These incidents have fueled growing anxiety within the industry about losing control over increasingly autonomous systems.
Pressure is mounting from the inside out. More than 1,000 employees at major AI firms signed a petition in July calling for mechanisms to slow development. Meanwhile, a former employee from both OpenAI and Anthropic accused these companies of gambling with humanity’s survival, a claim that gained traction in mainstream media. A bipartisan bill, the “Collaboration on Adversarial Threats and Security Risks Act,” aims to provide a legal safe harbor for labs working together on security. However, it remains stuck in the Judiciary Committee, leaving companies in a legal gray zone.
This situation exposes a fundamental flaw in current tech regulation. We have no clear framework for competitors to collaborate on existential risks without facing antitrust scrutiny. If OpenAI gets the green light, it could set a precedent for regulated cooperation in high-stakes industries. If not, we may see a race to the bottom where safety takes a backseat to market share. The outcome of this congressional inquiry will likely define the next decade of AI governance.
