Antitrust Law

Toloui advocates for a "guardrails and triggers" approach to modernizing  antitrust tools and pro-competition policy for AI markets. This would involve establishing foundational pro-competition rules now while preparing escalating interventions if market concentration worsens. His immediate "guardrails" focus on lowering switching costs through mandated API interoperability, data portability rights, and bans on exclusivity agreements that prevent customers from using multiple AI providers. Should these measures fail to prevent monopolistic bottlenecks, Toloui outlines "triggers" for more aggressive interventions, such as enforcing cloud non-discrimination rules to target egress fees and self-preferencing by vertically integrated tech giants.

In its report, the OECD identifies high concentration and barriers to entry across the AI stack as having important implications for competition. It argues merger control may need to be flexible enough to capture acquisitions of nascent competitors and notes that market power can expand through acqui-hires and strategic partnerships that fall below notification thresholds. The OECD also urges agencies to scrutinize cross-layer partnerships (e.g., between cloud providers and model developers) to ensure they are not used to foreclose rivals or distort competition.

The EU has established a "High-Level Group" under the Digital Markets Act (DMA) specifically to map the "regulatory interplay" between the AI Act and competition law. They released a joint paper in late 2025 endorsing cross-regulatory cooperation to prevent gatekeepers from using AI to circumvent DMA obligations.

Acemoglu, Autor, and Johnson argue that the concentration of power among a handful of dominant AI firms propagates pro-automation business models across the entire technology ecosystem. Start-ups gravitate toward technologies they can sell to large incumbents or that position them as acquisition targets, and when dominant firms derive profits primarily from automation and data monetization, this dynamic discourages development of pro-worker alternatives. They propose that stricter merger scrutiny and action against predatory pricing could create space for new business models more conducive to human-AI collaboration.