Skip to main content

Accountability

Being responsible for what an AI system does and being prepared to answer for it, you can't blame the algorithm and walk away.

The Simple Version

Being responsible for what an AI system does and being prepared to answer for it, you can't blame the algorithm and walk away.

Detailed Explanation

Accountability in AI is a multi-party obligation: providers are accountable for design and conformity; deployers are accountable for operation and oversight; governance boards are accountable for risk appetite and oversight adequacy. The EU AI Act distributes accountability across the supply chain through its provider, deployer, importer, and distributor obligations. The OECD AI Principles and ISO/IEC 42001 treat accountability as a core governance principle. Accountability requires traceability (to know who made decisions), transparency (to disclose information), and auditability (to allow verification).

Key Characteristics

  • Distributed across the AI supply chain: providers, deployers, importers, distributors
  • Requires traceability, transparency, and auditability as enabling conditions
  • Embedded in EU AI Act role obligations and OECD AI Principles
  • Cannot be contracted away, each actor remains accountable for their specific role

Why It Matters

Defining clear accountability assignments in AI governance frameworks prevents diffusion of responsibility, the single most common governance failure that leads to undetected AI harms and regulatory exposure.

Real-World Analogy

Like the clear chain of authority in a regulated financial institution, every decision has an identified accountable person whose name is on record, enabling regulators to ask 'who authorised this?' and receive a clear answer.

Common Misconceptions

  • Accountability for AI rests solely with the developer, it is distributed across all parties that deploy, import, or distribute AI systems.
  • Accountability is satisfied by having a signed policy, it requires demonstrated ability to identify, explain, and address AI outcomes, supported by documentation and oversight mechanisms.

Related Terms

Related Articles

Sources & Further Reading