Standards engagement is one of the least glamorous and often time-consuming areas of work relevant to international trade. However, development and adoption of standards — through rigorous, open, and consensus-driven processes — facilitate the frictionless flow of goods and services globally. They also lend credibility to business agreements, public procurement contracts, and international trade arrangements. That consensus model – used across the world by global, regional, national, and sectoral standards bodies alike – has been central to their acceptance. Until now, few have dared to touch this sensitive “third rail.”

Reports that Europe is supporting efforts to bypass consensus-based rules in its pursuit of AI standards has rippled far beyond Brussels. Presented as a procedural reset to better synchronize work with the standardization needs of the new AI Act, the decision of the European standardization organizations (ESOs) -- CEN and CENELEC -- to abandon their consensus-driven standards development is instead testing global confidence.

One immediate consequence has been that the standards group responsible for leading the global efforts for AI standards has hit the pause button on “parallel development,” the long-standing arrangement under which European and global bodies worked in tandem.

Why a pause? And why now?

The trigger lay in growing friction between EU regulatory imperatives and international standardization practices. On the one hand, the European Commission is eager to ensure that conformity assessments under the AI Act reflect EU legal definitions. On the other hand the Commission has expressed frustration at delays by the standards bodies to accelerate and complete the work that would provide the basis for these assessments.

To paraphrase Churchill, the wheels of standardization processes grind slowly; but they do grind exceedingly fine. Standards development processes were never intended nor designed to be at the beck and call of regulators. Instead, standards provide foundations – such as common terminology, agreed processes or guidelines, etc. – against which tests of conformity or compliance can be made, whether by regulators or in business agreements, contracts, etc.

The European Commission has instead demonstrated an appetite for tighter oversight of the standards process. Under pressure from the Commission, the body governing the European standards processes decided to its own rulebook to have a smaller -- but less accountable - group drive the work to completion. Many have argued that work should pause until it is clear how European Harmonized Standards will interface with global ones being developed by ISO/IEC JTC 1/SC 42, the main international AI committee.

While some – including those directly involved in the European efforts hinted that it is a politically motivated freeze that undermines Europe’s credibility as a neutral convener of global technical norms -- supporters of the short-cuts framed the CEN-CENELEC decision as necessary alignment with the AI Act’s risk-based obligations. However, the EU’s regulatory requirements – such as the conformity assessments required by the EU’s AI Act are not undermined by external, global, standards.

The European Commission seems determined that stakeholders should not consider this alternative approach – embracing existing global standards rather than developing their own. The EU’s own Standardization Regulation skirts around the issue by making only nodding references to the existence and value of international standards whilst rarely leveraging this body of existing work. The Commission has often used this apparent leeway to plow ahead with developing European standards, for which the European standards bodies receive financial recompense, rather than determine what standards already exist or are under development globally that would satisfy each legal requirement, rendering unnecessary the need to develop new, European only standards.

European Fallout

Within Europe, the decision risks fracturing the standardization community. National standards experts and bodies have expressed concern that halting work underway will waste years of technical effort and jeopardizes industry timelines for compliance. Others, backed the pause in parallel development, emphasizing sovereignty and the need to maintain coherence with EU law.

CEN-CENELEC’s committees have struggled to maintain momentum as working groups await clarification on whether resumed drafts will proceed as European-only deliverables or re-enter the ISO/IEC pipeline. For companies depending on harmonized standards to demonstrate AI Act conformity, uncertainty has turned into a business-planning headache.

Global Repercussions

The reverberations have been sharpest in the international arena. ISO and IEC officials privately warn that Europe’s withdrawal from consensus undermines the legitimacy of work already co-developed across continents. In the United States, stakeholders supported the pause in collaboration, citing as evidence that the EU is prioritizing regulatory sovereignty -- something the EU does vocally promote -- over interoperability, and thus complicating efforts to align trustworthy AI metrics.

The UK, post-Brexit, faces a dilemma: it remains within the European standards system, and is thus obliged to implement CE-CENELEC standards, and wants to thus maintain EU market access but is also seen as a major European player in the global standardization system.

Standardization agencies in Asian countries see an opportunity in the current vacuum that they can fill by accelerating their own frameworks — potentially redefining the global center of gravity for AI safety norms. I have argued in other fora that China would sometimes like to mimic the European standards strategy, using the same arguments used by the EU in order to defend and promote Chinese standards.

Commercial implications of any delays in this area are likely to be immediate. Multinationals implementing AI risk-management systems across jurisdictions now face diverging benchmarks, while startups risk being squeezed between “EU-only” and “global consensus” expectations.

The Strategic Tension: Regulation vs. Standards

At its heart, the crisis reflects a long-running philosophical divide. Regulation seeks control through law; standardization seeks consensus through practice. Creating standards whose sole objective is to implement regulation is a recipe for disaster, encouraging walled gardens around markets rather than being a means for different markets to interoperate. If, in addition, those standards are developed without an open, consensus-driven process, they could run counter to the World Trade Organization guidelines to avoid so-called “Non-Tariff Barriers to Trade” (TBT).

As the EU expands its regulatory perimeter – through the AI Act, Data Act, and Cyber Resilience Act – and ups the ante with its demands for so-called “digital sovereignty,” the space for voluntary, industry-led standardization is narrowing. The current pause in cooperation between European and global standardization efforts is symptomatic of a broader question: can Europe remain both a rule-maker and a consensus builder, or must it choose?

Paths Forward

As the European group developing European AI standards meets this week to determine its response to its global equivalent, many national experts are torn: many think that the Europeans already get “two bites at the cherry” – their national bodies having a vote in both the global and European groups; but on the other hand non-European national standards bodies do not get a vote on European standards that will affect many of the industries and stakeholders that they represent and who want to remain active in the European market.

Three scenarios seem plausible. Firstly, reconciliation: once conformity-assessment rules are finalized, the global standards bodies re-engage with their European counterparts, restoring credibility and joint authorship. Secondly, fragmentation: Europe proceeds with a distinct suite of “EU Harmonized AI Standards,” effectively creating parallel regimes. Or thirdly, “hybrid realignment:” limited cooperation continues, but under tighter Commission supervision and reduced openness to non-EU stakeholders.

Stakeholders will be watching CEN-CENELEC this week and in the coming months looking for signs from the result of global standards bodies ballots as well as official Commission guidance on standardization mandates.

Conclusion: A Pause That Speaks Volumes

Europe’s approach was framed as a safeguard for legal consistency, but its implications are geopolitical: it tests the durability of international trust in Europe’s ability to lead collaboratively and exposes the fragility of the global standards ecosystem just as AI governance becomes a matter of strategic competition.

However, based on public comments of the actors active in national, European, and global standards initiatives, there is a sense of a tectonic shift that is building tension between regulators and standard developers.

If consensus falters here, it may not just slow AI standardization — it could mark a turning point in the balance between regulatory sovereignty and the cooperative spirit that has long underpinned global technology governance.

If you or your organization have questions or are interested in standards engagement, please contact Peter at Venable.

‍