Amy Klobuchar

Take Google Maps. It is integrated with Google Flights and Google Hotel Ads, giving Google an advantage in search results over other flight and hotel aggregators like Expedia and Booking.com. However, the integration also fosters seamless search and booking experiences, which benefits users and pressures competing aggregators to offer greater value. Though currently legal in the United States, the DMA bans this integration with a blanket “self-preferencing” rule. The result? Hotel and flight bookings in Europe take far more clicks, costing consumers nearly $4 million in quantified wasted time yearly. The degraded user experience also drove a 30% loss in web traffic and a 36% loss in direct bookings on Google Hotel Ads, harming smaller hotels reliant on the service.

Instead of learning from Europe’s mistakes, AICOA would import them. Like the DMA, it targets various product integrations deemed to disadvantage the digital platform’s competitors, including cross-use of data between services, self-preferencing, and restricting competitors’ use of certain platform features.

AICOA’s proponents argue it maintains antitrust law’s “standards”-based flexibility. Unlike the DMA’s blanket bans, which are enforced by regulators with only narrow exceptions, competition enforcers must take firms to court for AICOA violations. Defendants can argue the practice does not “materially harm” competition, or is necessary for user privacy or security.

However, these differences are a veneer. “Materially harm competition” is a vague term, undefined in prior cases. AICOA describes it as “any actual or reasonable risk of lessening of competition … that is more than a de minimis amount.” This low bar prevents a holistic weighing of pro- and anticompetitive effects, advantaging competitors seeking a shield from competition rather than helping consumers. Affirmative defenses must also be proven by “clear and convincing evidence,” which is hard when the impacts of a novel or new technology, product integration, or business practice may not yet be clear, or when a counterfactual without the integration cannot be reconstructed. By creating substantial uncertainty and litigation risk, both AICOA and the DMA raise compliance costs and risk chilling experimentation and innovation that benefits users and small businesses alike.

RURAL AMERICA WAS TOLD ‘LEARN TO CODE’ — NOW IT MUST POWER AI

These harms aren’t theoretical. Uncertainties about what DMA compliance requires and how to balance compliance with privacy and security caused delays of up to a year before cutting-edge AI features including Apple Intelligence, auto-translate for AirPods, and Google search AI summaries reached EU users. A recent survey finds that EU and U.K. small tech firms forego approximately $109,000 to $375,000 per year due to delayed product and AI model launches. Almost 60% of U.K. and EU developers report launch delays, and over a third had to downgrade or cut features to comply. Studies estimate an AICOA-like policy could similarly cost American SMEs billions.

America’s world-leading tech sector emerged from vigorous competition and innovation under laws that generally favor commercial experimentation instead of punishing firms for their size, efficiency, or success. Europe has struggled to produce leading technologies and platforms despite its skilled and educated populace. Burdensome regulations that leave small and large tech firms in the dark without concomitant benefits don’t help. Just half of European startups actively use artificial intelligence today, compared to almost two-thirds in America. Importing their policy failures isn’t just unwise. It’s un-American.

Satya Marar is a research fellow at the Mercatus Center at George Mason University, specializing in competition, innovation, and governance. Andrew Liu is an intern at the Mercatus Center at George Mason University.