The study integrates two established theoretical frameworks to make sense of the divergence. The first is Fountain’s technology enactment framework, which holds that information technologies are not simply adopted by organizations but are actively shaped, or enacted, by the institutional structures that surround them. The second is Horowitz’s adoption-capacity theory, which argues that a state’s ability to absorb transformative military technology depends on the financial and organizational capital it can marshal. By combining these lenses, Parisini argues that money alone cannot explain why the United States and the European Union have followed such different trajectories, because organizational capital, the accumulated capacity of institutions to absorb and deploy new technology, is the critical determinant of effective AI implementation.

The comparison reveals two distinct governance models. The United States exhibits what the study calls competitive institutional appropriation, a capability-first approach characterized by concentrated investment and a race to translate AI research into operational military advantage. This model is anchored in a dense ecosystem of institutions, from the Department of Defense’s Data, Analytics, and Artificial Intelligence Adoption Strategy, published in 2023 under the banner of accelerating decision advantage, to executive orders from the White House on safe and trustworthy AI development and, later, on removing barriers to American leadership in artificial intelligence. The National Artificial Intelligence Initiative Act of 2020 further institutionalized federal AI spending, and analyses of US government AI budgets show sustained growth in defense-related investment.

The European Union, by contrast, demonstrates fragmented institutional absorption. Rather than a single centralized push, AI adoption in European defence is distributed across member states and a lattice of EU frameworks, including Permanent Structured Cooperation, or PESCO, established by Council Decision in 2017, the European Defence Fund, and the Artificial Intelligence Act of 2024. The study finds that strategic choices within these frameworks have tended to reinforce member state autonomy rather than pool it, meaning that national capitals such as Paris and Berlin continue to chart their own courses. France, for example, has pursued defence AI through a dedicated defence innovation agency and one of Europe’s fastest classified supercomputers for defence applications, while Germany’s approach has been shaped by its own strategic reorientation. The result is a patchwork rather than a unified program.

The financial asymmetry between these two models is stark and consequential. The study highlights significant disparities in financial intensity between US and European defence AI efforts, a gap rooted in decades of divergent defence expenditure across EU countries. That spending gap, the analysis argues, creates technological dependencies, with European armed forces and defence industries relying on American platforms, standards, and infrastructure for critical AI capabilities. This dependency relationship poses profound implications for European strategic autonomy, the long-standing ambition that Europe should be able to act in security matters without critical reliance on external powers. In a domain as sensitive as military AI, where data pipelines, model architectures, and decision-support systems can determine battlefield outcomes, dependence on another bloc’s technology is a strategic vulnerability.

European policymakers are not blind to the problem. The study points to emerging strategic initiatives, most notably the European Commission’s White Paper for European Defence, known as Readiness 2030, published in 2025, as an attempt to close the gap and rearm Europe’s technological base. The White Paper sits alongside continued financing decisions for the European Defence Fund, which channels research and development money into collaborative defence technology projects. Yet the study’s central caution is that these initiatives will not succeed on funding alone. Because organizational capital, the trained personnel, institutional routines, data infrastructure, and decision-making structures that allow organizations to actually use AI, develops slowly, a surge of money without institutional reform risks reproducing the fragmentation that currently limits European absorption capacity.

The theoretical stakes of the finding extend beyond Europe. By showing that divergent implementation pathways reflect institutional logics rather than technological imperatives, the study challenges a common assumption that AI will naturally converge toward similar governance forms everywhere. Institutional theory has long held that organizations in the same field tend to grow alike through isomorphic pressures, yet the EU and US cases show that where those pressures come from, whether competitive markets, regulatory frameworks, or strategic traditions, shapes fundamentally different outcomes. In the American case, governance has been framed around competitive advantage and national security leadership; in the European case, around regulation, ethical safeguards, and the preservation of national sovereignty within a collective framework. Both are responses to the same technology, but they enact it in incompatible ways.

These divergences carry direct consequences for defence cooperation between the allies. The study suggests that successful transatlantic collaboration on military AI requires attention to both technical and institutional compatibility, while actively mitigating dependency risks. Interoperability between an American capability-first ecosystem and a fragmented European one is not merely a question of common data standards or shared interfaces; it requires aligning institutional logics that currently pull in different directions. The research also situates the issue in a broader geopolitical context, noting that the United States’ dominance of global arms exports has grown while Russian exports have fallen, and that Ukraine has become the world’s biggest arms importer, trends that raise the stakes of who controls the next generation of defence technology.

The methodological approach is a comparative case study, a design suited to tracing how institutional structures and organizational factors shape technology adoption in complex political settings. By examining documentary evidence ranging from the US Department of Defense’s AI adoption strategy and White House executive orders to the European AI Act, the Readiness 2030 White Paper, and national defence AI strategies from France and Germany, the study builds a picture of two governance systems operating at different speeds and with different centers of gravity. The author reports no competing interests, and the article appears in Volume 5 of Global Public Policy and Governance, pages 114 to 138.

For readers tracking the future of military technology, the study’s message is clear: the AI race in defence will not be won by algorithms alone. Governance structures mediate which technologies get adopted, how quickly, and under whose control, and the gulf between American competitive appropriation and European fragmented absorption is widening into a structural dependency that money alone cannot close. Whether initiatives like Readiness 2030 can build the organizational capital Europe needs, and whether transatlantic institutions can find enough common ground to cooperate on the most consequential technology of the era, are questions that will define European strategic autonomy for decades to come.

Subject of Research: Comparative governance of artificial intelligence adoption in the EU and US defence sectors

Article Title: Governing artificial intelligence in the defence sector: a comparative analysis of EU and US institutions

Article References: Parisini, E. (2025). Governing artificial intelligence in the defence sector: a comparative analysis of EU and US institutions. Global Public Policy and Governance, 5(2), 114-138. https://doi.org/10.1007/s43508-025-00115-x

Image Credits: AI Generated

DOI: 10.1007/s43508-025-00115-x

Keywords: artificial intelligence, defence policy, European Union, United States, AI governance, military technology, strategic autonomy, institutional theory, European Defence Fund, AI Act, defence innovation, transatlantic cooperation

The authors are careful to stress that they are not advocating blind automation. Rights- and process-based constraints such as contestability, nondiscrimination, recourse, and accountable oversight can legitimately limit any decision system, whether human or machine. Their claim is instead one of symmetry. When audited evidence shows AI outperforming humans on endorsed criteria and the machine can satisfy core procedural safeguards, the justificatory burden shifts. At that point, they argue, asking only whether AI is safe enough to deploy leaves the harder question unasked: whether human control is safe enough to keep.

The argument rests on a substantial body of empirical research showing that human judgment is often less reliable than we assume. Decades of work dating back to Paul Meehl’s 1954 analysis of clinical versus statistical prediction, and the 2000 meta-analysis by William Grove and colleagues in Psychological Assessment, have repeatedly found that simple mechanical rules match or beat expert clinical judgment. Daniel Kahneman and colleagues’ work on noise documents how inconsistent human judgments are, even among trained professionals, while research on time-of-day effects shows that memory performance and other cognitive functions fluctuate with factors as mundane as the hour of the day. Ego depletion research further suggests that self-control, a resource central to careful decision-making, is limited and unreliable.

Aeon and Gruda trace the human gold standard fallacy to four recurring psychological drivers. The first is omission bias, the well-documented tendency to judge harms caused by action more harshly than equal or greater harms caused by inaction. When an algorithm makes a bad decision, the harm feels like something we did; when a human makes the same or worse decision, the harm is reframed as something that simply happened. Meta-analytic work by Sek Yeung, Timur Yay, and Gilad Feldman in 2022 confirmed robust omission-commission asymmetries in moral judgment, and philosophical analysis by Pinky Willemsen and Katia Reuter questions whether the omission effect is even coherent.

The second driver is agent-relative ethics, the philosophical position that what one is permitted to do differs from what one would prefer to see happen, making the identity of the decider morally salient. Research by David McNaughton and Piers Rawling on agent-relativity and the doing-happening distinction underpins this tradition, and the authors argue that such considerations make status-quo human authority feel morally safer even when it performs worse by our stated standards. The third driver they call intuitive exceptionalism, the deep-seated intuition that human minds occupy a special category, related to findings by Yoel Inbar and colleagues on people’s aversion to machines making moral decisions, and to taboo trade-off research by Philip Tetlock and colleagues showing that people react strongly when sacred values are weighed against mere utilities.

The fourth driver is the control premium: the value people place on retaining control for its own sake. Ellen Langer’s classic 1975 work on the illusion of control demonstrated that people systematically overestimate their influence over outcomes, and studies of decision-making under uncertainty show a propensity to under-delegate even when delegation would yield gains. Recent work by Anna Grundke on status threat suggests that outperforming machines evoke a threatened sense of human standing, while research by Nicolas Spatola and Antoine Normand documents the psychological costs of being compared to a superior artificial agent. Together, these four forces, the authors argue, make human authority feel better than machine authority even when evidence favors the machine.

The paper situates this argument within the current regulatory landscape, particularly the European Union’s Artificial Intelligence Act, which mandates human oversight for high-risk systems. The authors cite work by Jakob Laux on institutionalized distrust and democratic AI governance, by Leander Enqvist on what human oversight under the AI Act actually requires and by whom, and by Ben Green on the flaws of policies requiring human oversight of government algorithms. Legal scholarship on hybrid human-algorithmic decision-making by Tomas Enarsson, Leonhard Enqvist, and Marko Naarttijärvi, together with critiques of moral crumple zones by Madeleine Clare Elish, suggests that human-in-the-loop requirements often place humans in positions where they rubber-stamp automated outputs or absorb blame without exercising genuine control.

The empirical evidence on performance gaps is, in some domains, already striking. Generative AI has been shown to increase worker productivity in field experiments published in Science and the Quarterly Journal of Economics. Randomized controlled trials have found that AI tutoring outperforms in-class active learning in authentic educational settings. Large language models have been reported to outperform experts on challenging biology benchmarks, and work on AI-assisted prostate cancer MRI diagnosis examines whether domain experts can rely on AI appropriately. Studies of perceived fairness by Christoph Kern and colleagues and of public attitudes by Kevin Bansak and Elizabeth Paulson show that public acceptance of algorithmic decision-makers often depends on perceived performance, suggesting that intuitions may shift as evidence accumulates.

The authors also engage with the classic objection that delegating decisions to machines creates a responsibility gap. Andreas Matthias’s influential 2004 analysis of learning automata, subsequent work by Filip Santoni De Sio and Jero van den Hoven on meaningful human control, and Jilles Davidovic’s 2023 argument on the purpose of meaningful human control all grapple with the problem that machines may act in ways no one intended. Aeon and Gruda do not dismiss this concern, but they note that accountability research by Claude Novelli, Mariarosaria Taddeo, and Luciano Floridi, and by Atoosa Kasirzadeh and colleagues in the fairness, accountability, and transparency literature, suggests that institutional structures can attribute responsibility across human and machine actors. Meanwhile, legal scholars such as Diane Desierto and Frank Pasquale have long argued for due process rights in scored, automated societies, indicating that procedural protections need not presuppose human superiority.

The article’s central provocation is ultimately about epistemic fairness in policy design. If a society claims to care about accuracy, fairness, and welfare, then its governance frameworks should be symmetric in how they treat human and machine decision-makers: both should be audited, both should be subject to procedural constraints, and neither should enjoy an unearned presumption of legitimacy. The authors acknowledge that some decisions may remain legitimately reserved for humans on the basis of rights and dignity, but they insist that such reservations must be articulated and defended rather than smuggled in as defaults. Their closing challenge is likely to reverberate through AI governance debates for years: the burden of proof does not rest permanently on the side of automation. Once machines demonstrably serve our stated values better than we do, the burden falls on those who would keep control in human hands to explain why.

Subject of Research: The ethics of human versus AI decision-making authority and the burden of proof in AI governance

Article Title: When AI outperforms humans, who bears the burden of proof?

Article References: Aeon, B., & Gruda, D. (2026). When AI outperforms humans, who bears the burden of proof?. AI & SOCIETY. https://doi.org/10.1007/s00146-026-03375-1

Image Credits: AI Generated

DOI: 10.1007/s00146-026-03375-1

Keywords: artificial intelligence, AI governance, algorithm aversion, human oversight, decision-making, ethics of technology, omission bias, algorithmic accountability, AI Act, burden of proof, automation, moral psychology