KEY LEARNINGS
- AI accountability requires designating identifiable individuals who are answerable for system outcomes, ensuring that responsibility is not diffused across a machine or a team.
- The 'Many Hands Problem' in AI development makes traditional negligence hard to prove because harm often results from the interaction of many small decisions rather than a single error.
- Legal frameworks are shifting from a negligence standard toward strict liability for high-risk AI, meaning developers or deployers may be liable regardless of intent.
- The Three Lines Model provides a robust governance structure by separating risk ownership (First Line), risk oversight (Second Line), and independent assurance (Third Line).
- True accountability requires redress mechanisms, ensuring that individuals harmed by AI decisions have a clear path to challenge the outcome and receive compensation.
- 🌐European Commission: AI Liability Directive ProposalEC proposal for AI liability framework.
- 🌐NIST AI Risk Management Framework: Govern FunctionNIST guidance on governance and accountability.
- 📰IIA: The Three Lines ModelIIA guidance on the Three Lines Model.
- Bovens, M. (2007). Analysing and Assessing Accountability: A Conceptual Framework. European Law Journal.
- Nissenbaum, H. (1996). Accountability in a Computerized Society. Science and Engineering Ethics.
- Institute of Internal Auditors. (2020). The IIA's Three Lines Model: An Update of the Three Lines of Defense.
