Choice Screens Yield Mixed Antitrust Results After Two Years

On 28 April 2026, the European Commission released its first formal review of the Digital Markets Act and declared the two-year-old regulatory framework "fit for purpose." The phrase, tucked into the summary of a dense 180-page staff working document, was an institutional verdict on the most ambitious competition intervention in the digital economy since the Microsoft consent decree of 2002. The Commission noted a positive impact on consumers and competition, cited the emergence of alternative browser engines on iOS, and pointed to the choice screens that now greet every European who sets up a new smartphone or opens an updated operating system. The document, however, raised more questions than it answered: Were users actually switching? Were the screens designed well enough to overcome incumbency advantages that had been compounding for fifteen years? And, most pointedly, was the core remedy for self-preferencing simply redistributing default power rather than dismantling it?

The choice screen has become the default regulatory tool for the default problem. Platform self-preferencing, the practice by which a vertically integrated firm gives its own services preferential placement, ranking, or access on the infrastructure it controls, sits at the center of nearly every major competition action against Big Tech today. The theory of harm is straightforward: when Google places its own shopping results above organic listings, or when Apple pre-installs Safari as the only browser on a new iPhone, the platform exploits its gatekeeper position to steer demand toward its own offerings, foreclosing rivals regardless of quality. The remedy, in the European formulation, is equally straightforward: force the gatekeeper to present users with a neutral, prominently displayed menu of alternatives and let the market decide. What the Commission's two-year review did not resolve was whether that remedy, in practice, produces anything close to the competitive dynamics its architects envisioned.

The DMA's choice screen obligations took effect in March 2024 for the six designated gatekeepers, Alphabet, Amazon, Apple, ByteDance, Meta, and Microsoft, covering 22 core platform services. For Apple, the most visible obligation was the browser choice screen on iOS: upon first opening Safari after updating, European users were presented with a randomized list of alternative browsers, including Chrome, Firefox, Brave, DuckDuckGo, and Opera, and asked to select a default. For Google, the obligations extended to search engine choice screens on Android, a conceptually identical mechanism that had been tested years earlier under the EU's 2018 Android antitrust decision, which required Google to offer a search engine choice screen on new Android devices in the European Economic Area. Euronews reported that under the DMA, Europeans can choose their default browser, pay with rival apps, and message across platforms, a set of interoperability and choice mandates that the Commission treats as the regulatory floor.

Apple's response to the DMA has been a study in calibrated hostility. Kyle Andeer, Apple's vice president of products and regulatory law, used the occasion of the DMA's two-year review to deliver a broadside against the law's architecture. In a statement covered by AppleInsider, Andeer argued that the DMA's interoperability mandates created privacy and security risks that the Commission had failed to adequately weigh, and that the law's prescriptive approach was producing compliance costs disproportionate to any demonstrated consumer benefit. The company has also been required, since August 2024, to allow European users to delete pre-installed default apps including Safari, Mail, and the App Store, a concession Apple had resisted for years on the grounds that these apps were integral to the iOS experience. Apple implemented the changes but structured the deletion process in a way that critics argued was deliberately confusing, burying the option several layers deep in settings menus rather than surfacing it during device setup.

The tension between Apple and Brussels is not merely procedural. It reflects a deeper disagreement about what self-preferencing regulation is supposed to accomplish. The Commission's theory, embedded in Article 6(3) of the DMA, is that gatekeepers should not be permitted to "favor their own products or services in ranking and related indexing and crawling" relative to those of third parties. The prohibition is structural: it does not require a showing of consumer harm, only a showing of gatekeeper status and the existence of self-preferencing conduct. Apple's counter-argument, articulated most clearly in its public filings and executive statements, is that vertical integration of hardware, operating system, and default applications is a product design choice, not an antitrust violation, and that mandated choice screens degrade the user experience by substituting a regulatory process for the seamless integration customers paid for. In Apple's framing, the real self-preferencing problem is Google's, not its own, and the DMA has misidentified the relevant market.

Google, for its part, has been living with choice screens longer than any other gatekeeper. The 2018 Android remedy required Google to offer a search engine choice screen on new Android devices sold in the EEA. The initial implementation, a quarterly auction in which rival search engines bid for inclusion on the screen, was widely criticized by competitors as a pay-to-play scheme that monetized the remedy rather than enabling genuine competition. DuckDuckGo, in particular, argued that the auction mechanism forced smaller search engines to divert resources from product development to bidding, effectively turning the remedy into a new revenue line for Google. The Commission did not publicly endorse that critique, but the DMA's choice screen provisions notably omit any auction mechanism, requiring instead that alternatives be listed based on objective, transparent, and non-discriminatory criteria.

The question of whether choice screens work remains stubbornly empirical, and the data that exists is incomplete. A 2023 study by researchers at the University of Zurich, analyzing switching behavior in markets where Google's Android choice screen had been implemented, found that the screen's design elements, the number of options presented, the ordering of choices, and the language used to frame the decision, had substantial effects on outcomes, and that Google's market share remained above 95 percent in search across all measured jurisdictions. The Commission has not released its own switching data in the two-year DMA review, a silence that critics, including several members of the European Parliament, have cited as evidence that the choice screen experiment is underperforming. The Parliament, in its own resolution accompanying the review, called for "more rigorous enforcement" and noted that the DMA's goals would not be achieved if choice screens became a procedural checkbox rather than a genuine competitive forcing function.

Across the Atlantic, the choice screen question is arriving at a different procedural gate. In the United States, the Department of Justice's search antitrust case against Google, decided in August 2024 by Judge Amit Mehta of the U.S. District Court for the District of Columbia, produced a liability finding that Google had illegally maintained its monopoly in general search services and search text advertising through exclusionary distribution agreements, including the payments made to Apple and other device manufacturers to secure default search placement. The remedies phase, now underway in 2026, has become the central venue for the American version of the choice screen debate. Judge Mehta has signaled openness to behavioral remedies including a ban on default-search payments, a structural remedy that would functionally require choice screens on every new device sold in the United States. As the National Law Journal reported, the antitrust remedies order is taking effect after the judge denied Google's stay motion on May 8, 2026, writing that "there is no rule in this circuit that any disclosure of information is an irreparable harm sufficient to warrant a stay."

The procedural posture matters. In the EU, the DMA is an ex ante regulatory framework: the Commission designates gatekeepers, publishes obligations, and enforces without the need to prove harm in adjudicative proceedings. In the U.S., the Google remedies trial is an ex post competition-law proceeding in which the government must prove that a specific remedy is necessary to restore competition in a market where liability has already been established. The distinction, often elided in commentary, has real consequences for the design of remedies. An ex ante regulator can iterate on choice screen design through informal guidance and enforcement proceedings; an Article III judge crafting an equitable remedy must specify, in a single order, exactly what the screen must contain, how it must be presented, who must see it, and for how long. When the DOJ proposes a choice screen remedy in the Google case, it will have to survive the kind of exacting judicial scrutiny that the DMA's drafters, working in the more permissive environment of EU legislative process, never faced.

The political dimension has intensified in ways that complicate the legal analysis. CNBC reported that the European Commission has issued more than $7 billion in fines to U.S. Big Tech companies over the past two years, and the Trump administration has grown increasingly vocal in its opposition to what it characterizes as a discriminatory tax on American innovation. The DMA's choice screen mandates, along with the interoperability and data-sharing requirements imposed on gatekeepers, have become a flashpoint in transatlantic trade diplomacy. The U.S. Trade Representative has signaled that the administration views the DMA's cumulative compliance burden as a non-tariff trade barrier, a designation that would open a new front in the already fraught U.S.-EU commercial relationship. The self-preferencing question, in other words, is no longer being adjudicated solely in competition courtrooms; it is now being negotiated between trade representatives, and the outcome may depend as much on diplomatic leverage as on legal reasoning.

Meanwhile, the choice screen concept is expanding beyond its original search-and-browser frame. The European Commission spelled out in April 2026 how Google must share search engine data with third parties under the DMA, a measure that goes beyond surface-level choice to address the data asymmetries that make self-preferencing profitable in the first place. The preliminary findings propose that Google provide rival search engines with access to anonymized query logs and click data, enabling competitors to improve their ranking algorithms and reduce the quality gap that makes users gravitate back to Google even when presented with alternatives. This data-access remedy, which some competition economists have called a "deep choice screen," represents an acknowledgment that presenting users with alternatives is insufficient if the alternatives cannot compete on quality, and that quality, in algorithmic markets, is a function of data scale.

The data-sharing proposal also exposes a tension within the self-preferencing debate that choice screens alone cannot resolve. If the competitive problem is that Google's search engine benefits from two decades of accumulated query data that no entrant can replicate, then offering users a choice between Google and a materially inferior alternative is not a remedy; it is a ritual. The DMA's data-access provisions attempt to address this by requiring gatekeepers to share certain categories of data with business users, but the scope of the obligation remains contested. Google has argued that query-level data sharing creates privacy risks that the DMA does not adequately mitigate, an argument that echoes Apple's privacy-based objections to the DMA's interoperability mandates. In both cases, the gatekeeper is invoking privacy as a shield against a competition remedy, forcing regulators to weigh two statutory objectives, competition and data protection, that were not designed to be balanced against each other.

The judicial landscape in the United States adds another layer of uncertainty. Search Engine Journal noted that Google lost two major antitrust cases in 2024, one on search distribution and one on advertising technology, yet its stock appreciated by 65 percent over the same period, becoming the second most valuable company in the world. The market's apparent indifference to the legal losses reflects a judgment that the remedies are likely to be behavioral rather than structural, and that behavioral remedies in digital markets have historically been easy for incumbents to absorb. The 2018 Android choice screen did not materially alter Google's European search market share. The 2024 DMA choice screens, by the Commission's own muted acknowledgment, have not yet produced a measurable shift in browser or search engine market concentration. Investors appear to be betting that the U.S. remedies trial, however dramatic its liability findings, will produce a similar equilibrium: fines, compliance costs, and a margin of switching that does not threaten the core revenue streams.

That equilibrium is precisely what critics of the choice screen approach find most troubling. The self-preferencing problem, in this view, is not merely that gatekeepers steer users toward their own services. It is that gatekeepers have constructed ecosystems in which the default is sticky enough to survive any choice intervention short of structural separation. A user who selects DuckDuckGo from a choice screen but continues to use Google Maps, Gmail, and Google Calendar, all of which default to Google Search for web queries initiated within the app, has not escaped Google's self-preferencing architecture; they have merely changed the entry point. The DMA's drafters understood this and included prohibitions on "bundling" and "cross-use of personal data" that go beyond simple choice screen mandates. But the enforcement of those provisions has been uneven, and the gatekeepers, particularly Apple and Google, have been aggressive in testing the boundaries of what the Commission will tolerate in the design of compliance solutions.

Apple's compliance with the DMA's browser choice requirement illustrates the dynamic. When the obligation took effect, Apple presented a choice screen that listed Safari alongside eleven alternative browsers in randomized order. But the screen appeared only once, during device setup or after the first Safari launch post-update, and it offered no comparative information about the browsers: no privacy ratings, no feature descriptions, no user reviews. A first-time smartphone user, confronted with twelve unfamiliar names, was effectively being asked to make a decision without the information necessary to make it meaningfully. Several alternative browser developers, including Mozilla and the makers of the privacy-focused browser Brave, filed formal complaints with the Commission arguing that Apple's implementation was designed to maximize the probability that users would select Safari, either by default or by confusion. The Commission has opened an investigation into those complaints, but no findings have been issued as of May 2026.

The European Parliament, in its resolution on the DMA's two-year review, called for what it termed "choice architecture regulation," a more prescriptive approach in which the Commission would specify not merely that a choice screen must exist but exactly how it must be designed, how many options it must include, what information must accompany each option, whether options must be presented in a consistent order across users, and how frequently the screen must be re-presented. The resolution is non-binding, but it signals legislative impatience with the Commission's hands-off approach to remedy design. If the Parliament's view prevails, the choice screen will evolve from a market mechanism into a regulatory instrument whose every pixel has been vetted by competition authorities. The irony, not lost on the gatekeepers, is that a regulated choice screen may be even less threatening to incumbency than a genuinely unpredictable one, since every regulated detail becomes an opportunity for compliance arbitrage.

The U.S. remedies proceeding will have to confront the same design questions without the benefit of the DMA's iterative enforcement model. Judge Mehta, or the D.C. Circuit on appeal, will need to specify what a lawful default setting looks like, whether payment for default placement is per se illegal or merely subject to a rule of reason, and whether the remedy should apply to all distribution agreements or only those involving Google. The DOJ's trial team has indicated that it will seek a ban on all default-search payments, a remedy that would affect not only Google's agreements with Apple and Android OEMs but also the substantial payments that Microsoft makes to device manufacturers to secure default placement for Bing. That remedy would, in effect, impose a choice screen on every new computing device sold in the United States, a far more sweeping intervention than anything the DMA has attempted.

The choice screen debate, at its core, is a debate about the limits of consumer sovereignty as a regulatory theory. The concept assumes that users, given a fair presentation of alternatives, will make decisions that discipline dominant firms and open markets to competition. Two decades of behavioral economics research, and two years of DMA enforcement experience, suggest that this assumption is fragile. Defaults are powerful not because users are lazy but because the cognitive cost of evaluating alternatives is real, and platforms have optimized their ecosystems to make that cost as high as possible for anyone who tries to leave. A choice screen that does not account for that reality, that presents options without context, that appears once and disappears, that is designed by the very firm it is supposed to constrain, may satisfy a statutory requirement without altering a market outcome. The question the DMA's architects, and Judge Mehta, and the U.S. Trade Representative, must all answer is whether they are designing remedies, or rituals.