Simon Willison
Directly Responsible Individuals (DRI)
agentic
What happened
Simon Willison discusses the concept of 'Directly Responsible Individuals' (DRI) in human organizations and how it applies to LLM agents. He argues that an AI agent should never be a DRI because accountability is a uniquely human trait, referencing IBM's 1979 slide stating that computers cannot be held accountable and thus must not make management decisions.
Why it matters
It establishes a clear boundary for agent autonomy in professional workflows: agents execute, but humans own the outcome.
The take
A healthy reminder of the limits of delegation. While agents can execute complex workflows, human-in-the-loop oversight remains necessary for accountability. This is a critical design principle for enterprise agent architectures.
Do this
Design agentic systems with explicit human-in-the-loop checkpoints where a human DRI signs off on high-impact actions.
Don't read this site daily. Get it in your inbox.
The daily brief and Sunday deep dive — distilled, scored, and opinionated. For builders only.