Part 1 of the 2025 U.S. Tech Leadership and Regulation Series
The win by the Department of Justice (DOJ) in the liability phase of the antitrust trial against Google search comes at a critical historical moment when U.S. tech and economic leadership is facing increasing risks and global threats. However, despite these risks, the Biden DOJ sought to leverage the limited liability ruling to propose “remedies” that amount to radical tech regulations. This includes a prohibition on Google investing in rival large language models (LLM) or query-based artificial intelligence (AI) companies and startups. Moreover, remarks by leaders of DOJ Antitrust and the FTC under the Biden Administration indicating that they are targeting tech and AI, such as then-FTC Chair Lina Khan’s speech at the January 25, 2024 FTC Tech Summit indicating that she viewed antitrust enforcers’ role in response to the “the rapid deployment of artificial intelligence tools” is to address “basic questions of power and governance” through “policy choices we make now.” The FTC and DOJ Antitrust pursuit of similar cases against multiple leading tech platforms suggests that antitrust law may be used in ways that may critically undermine U.S. competitiveness and the immense benefits that the U.S. is poised to gain from its leadership in AI.
The AI “remedy” proposed by the DOJ in the Google search case is a paradigmatic attempt to impose regulations far afield from the specific issues raised during the trial and extending well beyond the case’s legal and economic foundation. This sets a poor precedent for future cases that could lead to de facto state planning via antitrust enforcement. More concretely, the AI remedy would harm consumers and restrict consumer choice, directly slow AI innovation and adoption, gravely undermine U.S. leadership in AI, hand foreign competitors a significant advantage, and jeopardize U.S. national security and economic interests.
Overreach from the Liability Phase Sets a Poor Precedent
The liability phase of the DOJ’s antitrust case primarily focused on Google’s practices in the search market. The DOJ argued that certain Google agreements with device manufacturers, such as Apple, and the default status on browsers that such agreements included, unfairly excluded rivals and entrenched a purported Google monopoly. Yet, there were no allegations, theories, or any evidence presented during this phase on Google’s investments in AI technologies, much less on acquisitions or investments in rival AI firms. In fact, nothing in the DOJ case suggested that AI is anything but a competitive market. This is unsurprising, as recent large-scale investments into numerous AI startups indicate that the AI segment is intensely dynamic and competitive, as is explored in more detail below.
This disconnect exposes how wholly inadequate are the proposed remedies. Antitrust remedies are helpful when they are tailored to address specific harms, they fail when they go beyond the violations to impose punitive measures or advance unrelated market regulations. Prohibiting Google from participating in acquisitions or investments in AI companies introduces a punitive element that goes beyond addressing the search-related issues adjudicated in court. It essentially implies that the DOJ can pick any other market to regulate independent of any finding of wrongdoing. Such a remedy could set a dangerous precedent where antitrust remedies become untethered from the specific competitive dynamics at issue in a case. That could amount to de facto state planning, or state selection of winners and losers in the tech sector, via antitrust enforcement activity.
Google Has Been a Key Driver of the AI Revolution
Requiring Google to effectively divest from AI would be shocking, especially as a remedy imposed in a case whose liability phase involved no discussion of any issues with competition in AI. Google has long been at the forefront of AI, investing heavily in research, talent, and infrastructure, including through acquisitions. In 2014 it acquired the London-based startup DeepMind, bringing in cutting-edge expertise that produced landmark achievements. For example, DeepMind’s AlphaGo became the first AI to defeat a world champion at Go in 2016, a feat that demonstrated how deep learning could tackle problems once thought impossible for machines.
DeepMind followed with AlphaFold, which in 2020 solved the 50-year “protein folding” grand challenge – predicting how proteins fold in 3D – a breakthrough hailed as a major scientific milestone. Google made this advancement freely available by releasing a database of 200 million protein structures predicted by AlphaFold, a resource that over a million researchers have used to accelerate vaccine development, drug discovery, and other innovations.
Beyond DeepMind, Google’s own AI labs (formerly Google Brain, now part of Google DeepMind) have driven fundamental research. Notably, Google introduced TensorFlow, an open-source machine learning library that has become one of the world’s most popular AI frameworks – empowering developers everywhere and speeding up AI development across industry and academia. Two Google-affiliated AI pioneers won the 2024 Nobel Prize in Chemistry, while another Google pioneer in machine learning won the 2024 Nobel Prize in Physics.
Google has also openly published AI research. For instance, the Transformer architecture that underpins modern natural language processing was developed at Google. Google has shared numerous tools and datasets, reflecting a commitment to public AI advancement. These investments have translated into consumer-facing AI products as well. Google’s large language model efforts (built on its research advances like Transformers and the LaMDA model) led to Google Gemini, a conversational AI assistant introduced under the name “Bard” in 2023. Gemini operates in over 40 languages and is integrated into services like Gmail, Docs, and Maps to help users with everyday tasks – from trip planning to summarizing emails.
In sum, Google’s investments – from fundamental research to open-source tools and real-world applications – have made it a pillar of AI innovation, contributing immensely to the field’s progress.
AI Is Critically Important for the U.S. Economy
Imposing an AI-focused remedy would be difficult to justify in any event, as artificial intelligence now represents one of the most dynamic and competitive sectors of the American economy, poised to drive innovation and productivity across nearly every industry. U.S. tech companies, from national champions like Google and Microsoft to a vibrant ecosystem of startups, are racing to develop AI solutions in areas such as healthcare, finance, transportation, and manufacturing. This intense competition is fueling rapid technological improvements and new business models. Venture capital investment in AI has skyrocketed – totaling roughly $290 billion in the U.S. over the past five years–reflecting the expectation that AI will unlock enormous economic value, including through AI startups.
Indeed, economists see AI as a general-purpose technology poised to boost growth on par with past breakthroughs like electricity or the internet. Goldman Sachs analysts project that widespread AI adoption (particularly advances like generative AI) could raise U.S. annual GDP growth by about 0.4 percentage points through the 2030s. Other estimates are even more bullish, suggesting AI could increase U.S. GDP growth by 0.5% to 1.5% per year (adding an estimated $1.2–3.8 trillion to the economy over the next decade) as companies deploy AI to automate tasks and create new products.
At the firm level, deploying AI often yields efficiency gains – from AI applications that boost software developers’ output to AI customer-service agents that allow human reps to handle more calls per hour. Across the economy these productivity improvements cumulate into higher overall output and can spur the creation of entirely new industries. For example, advancements in AI have accelerated the development of autonomous vehicles, personalized medicine, and other nascent sectors that promise future growth. In short, AI has become a key driver of American innovation and productivity, with U.S. companies leading in both cutting-edge research and commercialization – a position that contributes significantly to economic dynamism and future growth prospects, provided that AI innovation and adoption continue apace.
AI Is a General Purpose Technology at the Center of a Global Race for Leadership
Overregulating companies like Google could hobble U.S. competitiveness in the global AI race. It’s important to recognize that AI is much broader than just chatbots or content generators – it underpins critical advances in healthcare diagnostics, logistics, cybersecurity, national defense, and more. In effect, leadership in AI is increasingly synonymous with both economic leadership and geopolitical influence.
The United States and China are openly vying for primacy in AI innovation, given its potential to reshape the global economic order and confer strategic advantages. In this context, overly restrictive regulations on American tech firms could become self-defeating. If Google and other targeted U.S. companies are hamstrung by blanket rules or prohibitions – for instance, limits on AI investment and collaboration, on deploying advanced models, or on integrating AI features into their platforms – it would slow the pace of domestic innovation and implementation.
Such constraints would come at a time when China’s tech sector is accelerating: Chinese AI labs and companies (often with substantial state support) have made striking progress, exemplified by the recent emergence of China’s DeepSeek models that reportedly approach leading U.S. AI models’ performance while using only a fraction of the training compute power.
China is also aggressively pursuing AI adoption globally by offering its technologies to other countries and cultivating influence abroad. Analysts warn that if Washington overreaches with broad AI restrictions, it would “cede ground to China, creating an opening for Chinese AI firms to dominate” in many markets.
In other words, excess regulation at home could handicap America’s AI champions just as international competition intensifies, undermining U.S. leadership in a domain that will shape economic and security power for decades. Indeed, a Brookings Institution report cautions that the potential for overregulation threatens both U.S. economic innovation and national security leadership — particularly in the context of global competition with China. Heavy-handed restrictions on AI investment or deployment could weaken the very companies driving U.S. AI excellence and leave the nation flat-footed in one of the most pivotal technology races of our time.
Picking Winners and Losers in AI Is A Poor Idea
The proposed AI remedy in the Google search case singles out Google in a way that fundamentally alters the competitive landscape and arbitrarily picks winners and losers. Acquisitions and investments are legitimate strategies employed by firms to innovate, expand capabilities, and remain competitive in fast-evolving markets. Preventing Google from participating in these activities within the AI space would effectively hamstring one of the world’s leading technology companies, and would slow development of this technology, harming other businesses and innovation for consumers.
AI may be the next general purpose technology. Restricting Google’s activity in the AI space would be akin to restricting an early 20th century transportation manufacturer from investing in the automobile industry on the basis of objections to its leading position in the buggy whip market. Far from enhancing competition, such a limitation may severely undermine one of the clear leaders in AI investments and technologies. Moreover, knee-capping a leading U.S. AI firm that provides extensive open-source and public domain infrastructure, e.g., TensorFlow, Gemma AI models, and the very invention of the AI transformer, prevents startups and other firms from accessing fundamental AI infrastructure.
This restriction becomes even more problematic when considering that even if the DOJ/FTC were to succeed in cases against other targeted firms like Amazon, Apple, and Meta, many of Google’s competitors, both domestic and international, would face no such limitations. Direct competitors like Microsoft would remain free to acquire promising AI startups or invest in new technologies. International rivals, particularly from China, where firms like DeepSeek, Baidu, and Tencent are making aggressive AI investments with state support, would also operate without similar constraints. This asymmetry not only disadvantages Google but distorts the competitive playing field in AI innovation.
These distortions are likely to lead to slower innovation and slower AI adoption. In turn, consumers will get worse and fewer AI products, fewer AI features integrated with popular digital services, and less choice for consumers who won’t be able to select digital services integrated with Google AI. If the DOJ/FTC succeed in cases against other U.S. tech companies, consumers potentially face similar loss of choice with respect to those companies’ digital services and AI offerings as well.
U.S. Leadership in AI Would Suffer
The United States has long been a global leader in AI innovation, thanks in large part to the contributions of companies like Google, which typically invests tens of billions of dollars into research and development each year. However, maintaining this leadership position is far from guaranteed. Strategic rivals like China have made AI a national priority, with significant government support and rapidly advancing capabilities in areas like generative AI, facial recognition, and autonomous systems. Moreover, the race to deploy AI widely is a prerequisite to reap both the economic productivity and military superiority benefits of AI. Even if the U.S. maintains an edge in AI science and technology, a failure to lead in AI deployment relative to China could lead the U.S. to fall behind.
The DOJ’s proposed remedy risks undermining U.S. leadership by imposing unilateral restrictions on one of its most innovative and impactful firms. By sidelining Google from the acquisition and investment landscape and limiting how Google can integrate AI into its existing products, the DOJ could inadvertently tip the balance in favor of foreign competitors who face fewer regulatory constraints, and would directly slow U.S. efforts to increase AI adoption by limiting Google’s ability to add AI features to popular services. This dynamic would not only weaken the U.S. AI industry but also compromise not only productivity growth but also national security interests, given the strategic importance of AI technologies.
Furthermore, the prohibition on Google’s AI investments could deter other U.S. firms from pursuing aggressive innovation strategies. If companies perceive that success in emerging markets like AI will invite disproportionate regulatory restrictions, they may scale back their ambitions. Such a chilling effect would have long-term consequences for the competitiveness of the U.S. technology sector.
Remedies Should Address the Root Causes of Harm
Effective antitrust remedies must address the specific harms identified during the liability phase while minimizing collateral damage to competition, innovation, and consumer welfare. The DOJ’s proposed prohibition on Google’s AI acquisitions fails this test on multiple fronts, even if one accepts the questionable conclusions of the court in the liability phase.
A more effective, less costly, and less harmful approach would limit remedies to ensuring competition in the search market—the core issue of the case.
Such remedies would directly address the alleged anticompetitive behaviors without imposing sweeping restrictions on unrelated aspects of Google’s business. Importantly, it would avoid targeting Google’s broader innovation activities in markets like AI, where no direct harm has been demonstrated.
Conclusion
The DOJ’s proposed remedy to prohibit Google from acquiring or investing in rival query-based AI companies is a dramatic overreach that risks unintended consequences far beyond the scope of the case. By conflating concerns about search dominance with restrictions on AI innovation, the remedy unfairly penalizes Google, harms consumers, and jeopardizes U.S. leadership in a critical technological frontier.
Antitrust enforcement plays a role in preserving competition and preventing abuses of market power. However, it must be executed with precision and foresight, particularly in fast-evolving industries like technology. Remedies that stray from the core issues of a case and impose sweeping constraints on innovation not only risk harming the targeted firm but also undermine broader economic and strategic interests.
The DOJ should reconsider its approach and focus on remedies that address the specific competitive harms identified in the trial. Only by doing so can we ensure a fair, competitive, and innovative marketplace that benefits consumers, businesses, and the broader economy.
