What happened
IMDA Singapore published this discussion paper in May 2026, consolidating the conclusions of a working group of over 20 members of Singapore's legal community drawn from government, academia, and industry. The paper examines how civil liability should be allocated when AI agents act autonomously, use tools, interact with third parties, and cause harm. Three structural challenges are identified: first, that 'autonomous agents that can make decisions with reduced human involvement can diffuse accountability for outcomes'; second, that existing tort and contract law can handle many cases but faces 'significant practical challenges' of attribution across multi-actor agentic value chains; and third, that 'the unpredictability of agentic AI was also seen as a significant challenge, especially in scenarios where all parties took relevant safeguards, but the agent still caused harm in an unexpected way.' The paper stress-tests both fault-based (negligence) and strict liability regimes through a hypothetical of a computer-use agent that deviated from instructions and caused third-party losses, and identifies three areas for further study: value-chain responsibility allocation, protecting low-bargaining-power actors, and attribution of unforeseeable agent actions.
Why it matters
Any enterprise deploying AI agents — particularly in customer-facing or third-party-interacting roles — faces unresolved legal exposure that this paper maps with unusual precision; legal, compliance, and product teams need to understand the liability landscape before agentic deployments scale.
Action needed
Share with your general counsel and AI product leads; use the paper's hypothetical computer-use agent scenario as a template to assess your organisation's current agentic deployments against both negligence and strict-liability frameworks.