Why Industrial-Era Regulation Cannot Govern AI—and What Should Replace It

Traditional regulation is failing AI not because regulators lack will, but because the model they inherited was built for a slower economy. The rules that governed factories, railroads, and broadcast licenses assumed that technologies change at a pace legislatures can track. AI does not. The emerging consensus across policy institutions, standards bodies, and international scientific efforts is that effective AI governance requires two complementary shifts: agile, behavior-focused oversight that moves at technology speed, and technical standards that translate broad principles into auditable practice—developed together by governments, industry, and researchers rather than imposed from a single capital.

The Governance Gap: Why Traditional Regulation Fails AI

The core mismatch is temporal and structural. Writing for the Brookings Institution, a policy analysis of AI regulation traces the dominant regulatory philosophy to Frederick W. Taylor, whose doctrine of "enforced standardization of methods" worked in the industrial era largely because innovation and adoption moved slowly enough for detailed, uniform rules to remain relevant. Congress, the analysis argues, built oversight institutions on those precepts, producing regulatory micromanagement that cannot keep pace with digital-era change.

The second failure is one of fit. Because AI is a multi-faceted capability rather than a single product, "one-size-fits-all" regulation, as the Brookings analysis puts it, "will over-regulate in some instances and under-regulate in others." AI in a video game and AI that could threaten critical-infrastructure security have different effects and warrant different treatment—yet a uniform rulebook treats them identically. The International AI Safety Report, a scientific series created after the 2023 Bletchley Park summit to support internationally shared understanding of advanced AI risks, exists precisely because capabilities are advancing faster than national strategies can be written. The 2026 Singapore Consensus on global AI safety research priorities makes the same diagnosis from the research side: as frontier systems grow more capable, the central challenge is ensuring evaluation and defense move ahead of offense.

Agile Governance: Adapting to AI's Dynamic Nature

The proposed alternative is not deregulation but a different regulatory architecture. The Brookings analysis points to the telecommunications standardization process—1G through 5G, with 6G underway—as proof that standards-like processes can keep pace with successive technology generations. But it identifies a critical gap: those standards address technical issues, not the behavioral issues that arise when technology is applied. Its proposed remedy is agile oversight that adopts a standards-like process to develop behavioral standards, expressed as codes of conduct, enforceable by a dedicated agency.

This design inverts the traditional sequence. Instead of legislating detailed rules that age quickly, the regulator sets high-level principles and lets iterative, consultative processes fill in the specifics—updating them as capabilities change. The agency's remit centers on AI itself rather than bolting it onto existing authorities, and its agility frees it from the Taylorist micromanagement that made sense for steam-era industry. The trade-off is real: principles-based regimes depend on enforcement mechanisms and institutional trust to prevent them from becoming voluntary suggestions.

The Role of Technical Standards in Operationalizing AI Ethics

Standards do the work that ethical principles cannot do alone: they convert abstract commitments into repeatable, auditable practice. UNESCO's analysis of standards as governance tools notes that clear, repeatable guidance helps organizations minimize errors and increase productivity, and—more consequentially—that standards can support compliance with the essential requirements of AI regulations. In the UK, where the government has designated a standard, its use usually confers a presumption of conformity with relevant aspects of UK regulation; UNESCO reports that harmonized standards will play a comparable role under the EU AI Act, which classifies AI systems by risk and imposes stringent requirements on high-risk ones.

The institutional machinery is already being built. NIST, the U.S. national measurement institute, leads and participates in developing technical standards—including international ones—that promote innovation and public trust in AI systems, and as federal AI standards coordinator it works across government and with industry to identify critical standards activities, strategies, and gaps. In the UK, the Alan Turing Institute, the British Standards Institution, and the National Physical Laboratory have jointly established the AI Standards Hub, supported by the UK government, with a mission to advance responsible AI by unlocking standards as governance tools and innovation enablers.

Contrasting AI Governance Models

The table below contrasts the failing paradigm with the two modern approaches, showing how each addresses pace, mechanism, and participation differently.

Read together, the columns explain the diagnosis: the traditional model fails on pace and fit, agile governance repairs the pace problem through behavioral codes, and standards supply the concrete, auditable layer that makes both credible.

Navigating the Global Landscape: Interoperability and Multi-stakeholder Approaches

No single jurisdiction can govern a general-purpose technology. The academic analysis by Esmat Zaidan and Imad Antoine Ibrahim in Humanities and Social Sciences Communications argues that corporate involvement is needed to ensure laws facilitate rather than hamper innovation and competitiveness—some companies, they note, raised concerns that the EU's AI law proposal could drive investment out of the European market—and that corporations have a leadership role in applying ethical AI practices and endorsing global frameworks. UNESCO similarly identifies the need for strategic direction to identify standards that support both domestic compliance and global regulatory interoperability.

Convergence is emerging from the bottom up. The 2026 Singapore Consensus observes that different jurisdictions are identifying similar governance principles, and that frontier labs and major platforms have independently implemented similar controls—though it cautions that principles inherited from established security engineering are more mature and interoperable than newer ones, with open problems remaining around multi-agent security and agentic supply-chain risks. Its conclusion is direct: agentic risk management requires defense-in-depth and a multi-stakeholder approach.

Toward a Hybrid Model: Combining Agility and Standards for Robust AI Governance

The two modern approaches are complements, not rivals. Agile oversight supplies the adaptive, enforceable behavioral layer that prescriptive statutes cannot; technical standards supply the repeatable, internationally negotiable layer that codes of conduct need to be auditable. The EU AI Act already leans on harmonized standards for its conformity machinery, while the UK pairs principles-based regulation with the AI Standards Hub—evidence that hybrid designs are moving from theory to implementation. The remaining risk is fragmentation: if jurisdictions develop non-interoperable standards, the governance gap reopens at the border.

Charting a Path Forward for AI Governance

Policymakers and industry leaders should prioritize the development of hybrid governance frameworks that integrate agile regulatory principles with globally interoperable technical standards. The measurable indicator of progress is the establishment of new international multi-stakeholder working groups focused on specific AI risk domains, leading to the publication of harmonized, voluntary technical standards or conformity assessment schemes. The precedent of successive mobile-network standards shows the process can keep pace with technology; whether AI governance institutions can replicate it across jurisdictions is the question the next round of standards publications will answer.

Sources

• The three challenges of AI regulation — Brookings
• Enabling AI governance and innovation through standards — UNESCO
• International AI Safety Report 2026
• The 2026 Singapore Consensus on Global AI Safety Research Priorities — Aisafetypriorities
• Super Intelligence Standards — NIST
• AI Governance in a Complex and Rapidly Changing Regulatory Landscape — Nature