How American AI Giants Are Taking Over the Market

SUMMARY

Paul Krugman interviews Lina Khan, former FTC chair, on the growing dominance of American AI companies and the antitrust and consumer protection challenges they pose. The discussion explores the risks of monopolization across the AI technology stack, data privacy, copyright issues, and the impact on competition and innovation.

MAIN POINTS

• The FTC analyzed the AI technology stack for potential monopolization risks and emerging gatekeepers. [1:00]
• Concerns were raised about Nvidia and hyperscalers acting as core gatekeepers in AI, and about data privacy abuses by major firms. [3:01]
• The FTC investigated interlocking directorates and opaque partnerships among major AI firms, drawing historical parallels to early antitrust cases. [6:02]
• Copyright and fair use issues emerged as AI models trained on copyrighted content, threatening incentives for content creators. [8:01]
• AI-driven search is reducing traffic to publishers by providing summaries instead of direct links, undermining traditional media business models. [11:01]
• The emergence of efficient Chinese AI models like DeepSeek highlights conflicts of interest and the dangers of entrenched American tech monopolies. [14:00]

DETAILED ANALYSIS

Lina Khan, former chair of the Federal Trade Commission, outlined the agency’s multifaceted approach to regulating artificial intelligence, focusing on both competition and consumer protection. The FTC examined the entire AI technology stack, from hardware chips through cloud infrastructure to models and applications, to identify where monopolistic bottlenecks or gatekeepers might emerge. Drawing on lessons from earlier digital markets, Khan emphasized the risk that network effects and economic properties could allow a single layer—such as cloud computing or chip manufacturing—to dominate, potentially distorting the entire market and stifling innovation.

During her tenure, the FTC received reports from market participants about Nvidia’s growing role as a gatekeeper in AI hardware, as well as concerns about hyperscalers—large firms providing cloud computing—gaining privileged access across the AI supply chain. These developments raised alarms about the potential for entrenched players to pick winners and losers, undermining open competition. The agency also monitored consumer protection issues, particularly the hunger for personal data to train AI models.

Khan cited examples where companies like Google altered terms of service to allow the use of users’ emails and documents for AI training, often without adequate transparency or consent, raising legal questions about retroactive changes and data privacy.

Another area of focus was the proliferation of AI-driven scams, such as voice cloning fraud, where malicious actors use AI to impersonate individuals and commit financial fraud. The FTC sought to clarify that existing antitrust, data privacy, and consumer protection laws remain fully applicable to new AI technologies, countering industry arguments that new tech invalidates established legal frameworks. The agency also scrutinized interlocking directorates and complex partnerships among leading AI firms—such as those between Microsoft and OpenAI, or Google and Amazon—to determine whether these relationships granted outsized control to incumbents, echoing the historical origins of antitrust law in the era of corporate trusts like Standard Oil.

Copyright emerged as a contentious legal issue, with AI models frequently trained on copyrighted works without compensation or consent from creators. This practice has led to lawsuits and debates over fair use, as authors, artists, and journalists found their work used to generate competing AI content. The long-term implications for the sustainability of newsrooms and independent content production are significant, as AI-generated summaries increasingly replace direct links to original reporting, depriving publishers of essential web traffic and revenue.

This trend, already visible in Google’s privileging of its own vertical products in search results, has accelerated with the adoption of generative AI, threatening the financial viability of traditional media.

Khan’s recent op-ed, discussed in the interview, argued that the most transformative innovations have historically arisen from competitive markets, not from entrenched monopolies. She warned against allowing dominant tech companies to dictate the trajectory of AI development, especially as the emergence of efficient Chinese models like DeepSeek demonstrates the risks of complacency. The discussion concluded by drawing parallels to other industries, such as aerospace, where lack of competition led to negative outcomes, reinforcing the need for robust antitrust enforcement to ensure innovation and public benefit in the AI era.

LINKS

• Transcript of Paul Krugman's conversation with Lina Khan on AI and antitrust.
• Subscription page for Paul Krugman's Substack newsletter.
• Paul Krugman's profile on Bluesky social network.

KEYWORDS