If your strategic view of global AI regulatory authority assumes that the US position as home to the leading AI developers and China's position as the second AI superpower naturally translate into global regulatory credibility, the 2025 Pew Spring Global Attitudes Survey inverts that ranking. Across 25 countries surveyed, a median of 53% said they trust the European Union to regulate AI effectively. The United States: 37%. China: 27%. The EU is now the most globally trusted AI regulatory authority, by a substantial margin. The observation that follows is consequential for organisations setting global AI compliance, market positioning, and regulatory engagement strategy.

Global trust to regulate AI effectively: EU vs US vs China, 2025

The three-way distribution of trust is consistent across most of the 25 surveyed countries:

  • The EU receives the highest trust ranking in most countries surveyed (53% median).
  • The US receives mid-range trust (37% median), substantially below the EU but above China.
  • China receives the lowest trust ranking in nearly every country surveyed (27% median).

One caveat matters: the median across 25 countries is a different metric than the average preferences in any single country. Within the EU, trust in the EU's regulatory ability varies: German and Dutch respondents trust the EU more than Greek and Italian respondents. The aggregate EU-trust position is built on widespread modest support rather than uniform strong support. The same is true of the US and China positions: the median figures hide significant within-country variation.

Within the United States specifically, views are evenly divided between trust (44%) and distrust (47%) of the US government's ability to regulate AI effectively. US respondents trust the EU at 43%. Within-country US sentiment matches the global pattern: Americans trust the EU's regulatory ability roughly as much as their own government's.

Three structural drivers explain the EU's global credibility lead.

The first driver: the EU AI Act and the "Brussels Effect." The EU AI Act, with its first measures effective February 2025 and general-purpose AI obligations effective August 2025, is the most comprehensive AI regulation enacted globally. The substantive framework covers risk-based classification, high-risk system requirements, general-purpose AI obligations, and prohibited applications, and is widely viewed as the most systematic approach to AI regulation in any major jurisdiction. The "Brussels Effect" of EU regulation shaping global practice has been documented across GDPR, DSA, and other digital regulation; the AI Act extends the pattern. Global respondents likely associate the EU's track record on comprehensive digital regulation with regulatory capability on AI.

The second driver: the US AI policy reversal of January 2025. The rescission of Executive Order 14110, the subsequent deregulatory pivot, and the December 2025 federal-state preemption framework have all been visible. The deregulatory direction may be viewed by global respondents as less protective of public interest than the previous framework. The 37% US trust median may reflect this perception. The pattern holds within the US too: only 31% of Americans trust their own government to regulate AI (Ipsos data), alongside the 37% global median for US trust.

The third driver: China's regulatory framework, despite being substantial, faces a credibility gap from global non-Chinese respondents. China has enacted significant AI regulation (generative AI service interim measures, AI-generated content labelling, deepfake regulations). The substantive content is meaningful. But global respondents from countries other than China appear to discount Chinese regulatory authority either because of broader scepticism toward Chinese institutions or because Chinese regulation is viewed as oriented toward state interests rather than public protection. The 27% global median for China is the lowest of the three, despite genuine regulatory activity.

A further caveat: the public trust rankings do not directly correspond to actual regulatory enforcement capability. The EU's strong public trust is partly aspirational: the AI Act enforcement is at early stages, and the operational compliance machinery is still being built. The US trust position is below its actual institutional regulatory capability. The China trust position is below its actual regulatory activity volume. Public trust is a different metric than enforcement effectiveness.

Three structural implications follow for organisations setting global AI regulatory strategy in 2026.

The first implication: aligning AI practice with EU AI Act requirements provides not just EU market access but also global credibility positioning. Organisations that meet EU AI Act requirements can credibly claim compliance with the most globally trusted regulatory framework. Organisations that meet only US federal requirements lack this credibility positioning. The market value of EU AI Act compliance extends beyond EU operations to global brand positioning.

The second implication: US-headquartered AI organisations face a structural credibility gap when operating internationally. The 37% global trust median for US government regulation translates into reduced credibility for AI products from US-based vendors when global customers consider regulatory compliance and trust. The credibility gap can be partially closed through organisational practice (visible commitment to AI governance, third-party audits, voluntary standards adoption) but the structural disadvantage relative to EU-based or EU-compliant vendors is real.

The third implication: the EU's global regulatory authority position has competitive implications for AI development location and partnership decisions. Some organisations are considering EU-based AI development hubs to leverage the regulatory credibility. The EU's compute infrastructure expansion (EuroHPC), regulatory clarity (AI Act), and trust positioning create a coherent strategic case for AI development presence in the EU even when the US offers more capital and frontier-model concentration.

A note on the data: the Pew Spring 2025 survey covers 25 countries, fewer than the 30-country Ipsos sample but provides directly comparable cross-country trust rankings. The methodology asks respondents to rate trust in three specific regulatory authorities (EU, US, China) rather than open-ended trust questions. The structure of the question may itself shape responses: respondents primed to consider three options may answer differently than those asked open-ended questions. The cross-country consistency of the EU-US-China ranking is robust to this caveat.

For organisations setting global AI strategy through 2028, planning needs to engage with the actual global regulatory credibility distribution rather than the assumption that US and Chinese AI development scale produces commensurate global regulatory credibility. The EU's regulatory authority position is now the most globally trusted. The strategic implications for compliance, market positioning, partnership decisions, and development location follow from this reality. Plans calibrated against the actual global trust distribution will be aligned. Plans operating from the assumption that US institutional position translates automatically to global credibility will be misaligned.

The 53-37-27 distribution across the EU, US, and China is one of the clearer signals in global AI public opinion data. The cross-country consistency, the multi-year stability, and the alignment with the EU's regulatory output all support reading this as a structural feature of the 2026 global AI regulatory environment rather than a transient pattern. Strategic plans through 2028 should treat the EU's regulatory credibility lead as a planning variable.