A surge of new AI rules and standards is fracturing interoperability: mismatched technical specs and divergent national regulations are raising compliance burdens and complicating cross-border deployment for firms and regulators.

Main Finding

AI governance is rapidly proliferating across hard laws (legislation/regulation) and soft laws (standards, certifications, principles), producing a fragmented and fast-moving landscape that raises compliance costs and market frictions. The report argues that combined technical and regulatory interoperability—aligned, adaptable standards, verification/certification, and coordinated roadmaps across public, private, and civil-society actors—is essential to reduce fragmentation, enable cross-border trade in AI services, lower barriers to entry, and preserve innovation while managing harms.

Key Points

• Definitions
  • Technical interoperability: connections between systems/services/platforms to enable data management, scalability, and user experience.
  • Regulatory interoperability: intentional overlap and consistency between rules across jurisdictions and governance instruments to facilitate trade and reduce regulatory arbitrage.
• Current landscape
  • Hard laws: Rapid emergence of national, subnational, and sectoral laws and executive actions (examples: EU AI Act; shifting US federal EOs and NIST mandates; US state laws and local rules such as Colorado SB 24-205 and NYC Local Law 144; Canada’s AIDA process and other initiatives).
  • Soft laws: Proliferation of standards (ISO/IEC 42001), certification programs, codes of conduct, and principle documents (e.g., OECD AI Principles).
• Problems identified
  • Fragmentation and jurisdictional divergence create uncertainty for firms, especially transnational ones (which standards/regulations to follow; which auditors to hire).
  • Rapid technological change risks obsolescence of static rules.
  • Limited existing attention to technical interoperability in hard law; much of technical spec work currently in soft-law fora or industry bodies.
  • Sector-specific vs. cross-cutting approaches generate coordination challenges.
• Lessons from other industries (case studies planned/used)
  • Learnings from nanotechnology (NanoDefine), environmental geodata (EU INSPIRE), telecommunications (historical evolution), and internet/web architecture (HTTP/IP): design adaptive governance, establish early measurement/definition standards, build trust via verification, and bridge technical-regulatory divides through collaboration.
• Proposed pathways (high level)
  • Public sector: incorporate international frameworks and standards into procurement and policy; create sandboxes, measurement tools, and roadmaps.
  • Technical: promote open software standards, sovereignty-compatible architectures, and certification/assurance layers.
  • Cross-cutting: build an economic/trade case for interoperability and broaden representation in standards-setting.

Data & Methods

• Research period: November 2024 – November 2025 (sections frozen as of October 2025). Funded by Google.Org (no author conflicts declared).
• Evidence and inputs:
  • Literature and policy review of hard and soft law instruments.
  • Use of existing datasets cited by authors (e.g., Faveri & Marchant datasets on 200+ US state and federal AI laws and 1,400+ international AI-related standards; OECD AI Policy Observatory; EAIDB Responsible AI Ecosystem Market Map).
  • Comparative case-study methodology drawing on four cross-industry cases (nanotechnology, environmental sustainability/INSPIRE, telecommunications, Internet/web architecture) to extract governance lessons.
  • Multi-stakeholder engagement via an advisory group, webinars, op-eds, and an in-person workshop.
• Analytical framework: qualitative mapping of governance instruments, identification of interoperability challenges, and development of sectoral roadmaps (public, private, NGO/CSO, and international standard-setter roadmaps) with phased implementation proposals.
• Limitations:
  • Fast-evolving policy and technology landscape; the report’s update cutoff (Oct 2025) may omit subsequent developments.
  • Emphasis on mapping and pathways rather than quantitative impact estimation.

Implications for AI Economics

• Transaction and compliance costs
  • Fragmentation raises firms’ compliance costs (multijurisdictional legal work, multiple audits/certifications) and slows product deployment across borders.
  • Interoperability (mutual recognition, harmonized standards, shared certification) can reduce these transaction costs and lower entry barriers, especially for SMEs and cross-border service providers.
• Market structure and competition
  • Standardization can generate strong network effects: common standards lower switching costs and can expand markets, but early proprietary dominance may entrench incumbents.
  • Open, multi-stakeholder standards and certification regimes can mitigate lock-in and encourage competition; closed or unilateral standards risk increasing concentration.
• Trade and cross-border services
  • Regulatory interoperability facilitates cross-border trade in AI goods and services by reducing non-tariff barriers and legal uncertainty (beneficial for digital trade and global AI supply chains).
  • Geopolitical fragmentation (incompatible standards or export controls) could fragment markets and raise costs of international market access.
• Innovation incentives and dynamic efficiency
  • Clear, harmonized frameworks reduce regulatory uncertainty and improve investment signals, supporting R&D and diffusion.
  • However, overly prescriptive technical rules can stifle innovation; adaptive governance and performance-based standards better balance safety and innovation.
• Information asymmetry and trust
  • Certification, assurance, and verification mechanisms help reduce information asymmetries between providers, purchasers, and regulators—facilitating market transactions (e.g., procurement decisions, insurance markets).
  • Public-sector procurement requirements aligned with interoperability goals can create demand-side incentives for compliant products and spur standard adoption.
• Policy levers to realize economic gains from interoperability
  • Promote convergence: mutual recognition agreements, incorporation of international standards into domestic regulation, and harmonized certification schemas.
  • Support open standards and reference architectures to lower entry costs and reduce lock-in.
  • Use procurement and grant funding to create market demand for interoperable, certified AI systems.
  • Invest in measurement, testing, and verification infrastructure (sandboxes, benchmarks) to signal compliance and reduce asymmetric information.
  • Design sovereignty-compatible approaches that balance data-flow needs for AI competitiveness with national security and privacy concerns.
• Potential trade-offs and risks
  • Rapid harmonization without broad representation risks exporting dominant jurisdictions’ preferences and increasing geopolitical tensions.
  • Standards capture by incumbents can entrench market power; governance must ensure broad participation and transparency.

Summary implication: Achieving regulatory and technical interoperability is economically valuable—by lowering costs, expanding markets, and enabling efficient diffusion of AI—provided governance is adaptive, inclusive, and paired with verification/certification systems that preserve competition and respect legitimate sovereignty and security concerns.