Definition
The delegation of accountability alongside the task, in systems where nobody both controls the behaviour and can be held responsible.
What is at stake
It stood on philosophy alone for twenty-two years. It now has causal evidence under it, which changes what you can say in a governance conversation.
Where the word came from
- The responsibility gapAndreas Matthias, 2004
- Blurred accountabilityKropp's AI-agent naming experiment, 2026, the first causal evidence under it
How episcope came to it
Andreas Matthias, 2004, a philosopher of machine ethics at Lingnan University in Hong Kong, in a peer-reviewed article in Ethics and Information Technology. His case was the learning automaton: a system whose behaviour its maker can no longer predict, so the maker cannot fairly be held responsible for it, and nobody else controls it.
What to watch for
Matthias wrote about learning machines in general, twenty years before chat assistants. The gap he describes is structural. Kropp's 2026 experiment with 1,261 managers, reported in Harvard Business Review, shows a second route to it, through the name given to the system, which needs no learning machine at all.
Papers and articles behind it
Matthias, A. (2004). The responsibility gap: Ascribing responsibility for the actions of learning automata. Ethics and Information Technology, 6(3), 175-183.
Kropp, M., Bedard, J., et al. (May 2026). Research: Why You Shouldn’t Treat AI Agents Like Employees. Harvard Business Review.