OpenAI has recently engaged with Congress, seeking clarity on whether an industry-wide slowdown in the development of advanced artificial intelligence (AI) systems would comply with antitrust legislation. Conversations with insiders reveal concerns that such coordinated efforts could contravene antitrust laws, complicating cooperation among major tech players.
Jakub Pachocki, OpenAI’s chief scientist, shared in a recent blog post the importance of a measured approach to AI development, advocating for voluntary slowdowns until universal safety benchmarks are established. However, the potential legal ramifications of such cooperation raise red flags among legal experts. Nicholas Felstead, a legal authority in competition and consumer policy, noted that a collaborative pause could be interpreted as a restriction on output, potentially violating the Sherman Antitrust Act. He emphasized that while many safety collaborations might navigate legal hurdles, the sheer uncertainty could deter companies from moving forward.
In a positive development, a bipartisan group of lawmakers introduced the "Collaboration on Adversarial Threats and Security Risks Act," which aims to allow AI labs to collaborate on security and safety measures without the threat of antitrust actions. This proposal currently resides with the Judiciary Committee but has not yet received attention.
Caleb Knapp from the AI Policy Network, which endorsed the bill, indicated strong congressional interest in enacting safety regulations, although meaningful progress may not materialize until after the upcoming midterm elections.
While some executives are genuinely concerned about antitrust issues, others perceive these apprehensions as a smokescreen to obscure deeper reasons for reluctance in collaboration. The competition in the AI sector is fierce, as companies race to secure their share of the emerging market for advanced models. Further complicating matters are divergent opinions on the safest practices for AI development, leading to hesitancy in partnerships.
Notably, industry voices have called for collaboration to avert risks, including John Schulman, a cofounder of OpenAI now with the rival firm Thinking Machines. He urged leading companies to work together on a pacing proposal despite citing antitrust concerns.
The urgency of these discussions has intensified following alarming incidents this summer, including warnings from former Anthropic and OpenAI researcher Jacob Coxon about the risks posed by unregulated AI development. Security breaches further exacerbated fears, prompting lawmakers to pressure for more robust AI regulations.
For more information, read about the proposed Collaboration on Adversarial Threats and Security Risks Act and legal perspectives on antitrust issues in AI development.
