Researchers have identified a critical gap in AI governance: while organizations have embraced ethical principles like fairness and transparency, they lack practical mechanisms to enforce them. A new study published in AI and Ethics suggests blockchain technology could serve as the missing infrastructure layer, enabling real-time monitoring and accountability throughout an AI system's development and deployment.

What Is the "Principles-to-Practice" Gap in AI Governance?

Over the past few years, governments, industry bodies, and civil society organizations have introduced numerous frameworks emphasizing ethical AI principles. However, these guidelines remain largely abstract. Organizations struggle to translate high-level values into concrete, enforceable practices within their systems. This disconnect has real consequences: failures in AI governance across sensitive domains like healthcare, manufacturing, finance, and public services can result in serious economic, social, and reputational damage.

The problem is particularly acute because modern AI systems are increasingly complex, opaque, and adaptive. Traditional compliance approaches, which typically focus on auditing after deployment, cannot keep pace with the speed at which AI systems evolve and make decisions. Researchers acknowledge that this principles-to-practice gap remains one of the central challenges of contemporary AI governance.

How Could Blockchain Serve as an AI Governance Infrastructure?

Rather than using blockchain solely as a tool for post-deployment audits or data management, the proposed framework conceptualizes it as a governance layer that operates throughout an AI system's entire lifecycle. This approach would enable continuous monitoring, validation, and accountability from initial development through deployment and beyond.

The framework identifies specific governance-relevant data that could be stored on a blockchain layer at each stage of an AI system's lifecycle:

• Development Phase: Documentation of design decisions, training data sources, model architecture choices, and ethical review processes that establish the foundation for responsible AI.
• Deployment Phase: Records of system performance metrics, user interactions, and real-world outcomes that enable ongoing monitoring and validation of AI behavior.
• Operational Phase: Continuous logs of decisions made by the AI system, user feedback, and any incidents or anomalies that require investigation or correction.
• Accountability Infrastructure: Immutable records that create clear chains of responsibility across organizations, supporting inter-organizational traceability and coordination.

By mapping AI system risks to blockchain's core characteristics, the framework offers a structured pathway for translating abstract governance principles into actionable, system-embedded practices. The approach supports not only ex-post auditing but also ongoing monitoring and validation, addressing the speed and complexity challenges that traditional governance mechanisms cannot handle.

Why Does This Matter for Organizations and Policymakers?

The research acknowledges that blockchain is not a panacea for AI governance. However, the framework provides practical guidance for organizations seeking operational mechanisms to govern AI systems throughout their lifecycle. This is particularly valuable for organizations of varying sizes and across different sectors, as the approach can be adapted to different organizational contexts and regulatory environments.

The study highlights key technical, organizational, and societal considerations that must be addressed for effective implementation. These include ensuring that governance mechanisms are compatible with existing organizational workflows, that data stored on blockchain layers remains secure and privacy-compliant, and that the governance infrastructure can scale as AI systems become more prevalent.

For policymakers, the framework offers a foundation for developing enforceable AI governance standards. Rather than relying solely on principles-based guidelines, regulators could mandate that organizations implement governance infrastructure that provides continuous transparency and accountability. This shift from abstract principles to embedded governance mechanisms could significantly strengthen AI oversight across industries.

What Are the Next Steps for Researchers and Practitioners?

The paper provides conceptual guidance but acknowledges that significant empirical and design-oriented research remains necessary. Future work will need to address technical implementation challenges, such as determining optimal data structures for blockchain governance layers and ensuring interoperability across different AI systems and organizations.

Practitioners seeking to operationalize AI governance can use the framework as a starting point for identifying which governance-relevant data should be tracked, how that data should be structured, and what organizational processes need to be redesigned to support continuous governance. The research emphasizes the importance of considering local socio-cultural, regulatory, and institutional contexts when developing AI governance approaches, recognizing that governance solutions cannot be one-size-fits-all.

The convergence of blockchain and AI governance represents a significant opportunity to move beyond the current fragmented landscape of conceptual guidelines. By embedding governance mechanisms directly into AI systems, organizations and regulators can create enforceable accountability structures that keep pace with the rapid evolution of artificial intelligence technology.