Glossary · EU AI Act: high-risk AI
Stop button or similar procedure (AI Act)
Also known as: AI Act stop button, Interrupting an AI system, Stop function (AI Act)
German: Stopptaste oder vergleichbares Verfahren (KI-Verordnung)
Under Article 14(4)(e) of the EU AI Act, the human oversight measures of a high-risk AI system must enable the people assigned to oversight to intervene in its operation or interrupt it through a “stop” button or a similar procedure that allows the system to come to a halt in a safe state.
- AI regulation
- EU
- Machinery
In one sentence
Art. 14(4)(e) AI Act: people overseeing a high-risk AI system must be able to interrupt it through a stop button or similar procedure, ending in a safe state.
Example
The operator of an AI-based route planner for autonomous mobile robots can stop the planner from the control room; the robots then finish their current move under the fleet manager and stand still.
Explained in context
This term is a station of a mini course
- When an AI agent crosses the line: detect, contain and restore operationsStation: The containment ladderStart the course
Context cards connect this term with others to answer one question. Also in British English and German.
- Containing an AI agent: pause, revoke, isolate, shut downHow do you stop an AI agent that has crossed its boundary — without making things worse, and without confusing it with stopping a machine?UKDeutsch
- When an AI agent crosses the line: detect, contain and restore operationsWhat has to be in place when an AI agent does something it isn’t allowed to do — in IT systems, IoT fleets, OT and on machines — and how is it detected, stopped and released again?UKDeutsch
How it applies
- Part of human oversight: Together with the abilities to understand the system, interpret its output, decide not to use it and override or reverse it (Art. 14(4)(a)–(d)), see human oversight.
- “Safe state” of the AI system: The halt must leave the AI system and what it controls in a safe state — no half-finished transactions, no orphaned commands. What that state is must be defined for each system.
- For AI agents: The agent kill switch is a common way to implement the interruption for agents; it must work without the agent's cooperation.
- When it applies: With the high-risk obligations — December 2, 2027 for Annex III systems and August 2, 2028 for Annex I Section A products (Digital Omnibus on AI).
Stop procedure (AI Act) vs. emergency stop (machinery)
The AI Act's stop interrupts an AI system. An emergency stop is a safety function of a machine under machinery law, designed to avert hazards to people with validated stopping behavior and a defined safe state. Where an AI function controls a machine, both exist side by side: the AI stop never replaces the emergency stop, and the emergency stop doesn't depend on the AI system.