A G7 Strategy for AI Competition and Consumer Rights

Abstract

The rapid evolution of foundation models poses significant challenges for global competition policy and artificial intelligence (AI) liability frameworks due to their high fixed training costs, low marginal deployment costs, and ability to perform diverse tasks after being trained on vast datasets. Unregulated AI markets risk supply chain vulnerabilities, increasing costs, and perpetuated biases as dominant firms consolidate control, limit competition, and shape outputs in ways that harm competition and consumer rights. G7 competition authorities should increase scrutiny of AI mergers, acquisitions, and exclusive partnerships, enforce fair and equal access to GPUs and cloud resources—meaning access that is impartial and available to all companies on the same terms—mandate interoperability, and prevent exclusive deals that stifle competition in AI infrastructure. G7 economies should adopt AI liability frameworks that shift the burden of proof for AI-related harms onto AI companies, ensuring consumer protection by holding companies fully liable for high-risk AI harms unless safeguards are proven.