Article 14 of the EU AI Act: Human Oversight
Article 14 is the safeguard against automation bias: high-risk AI must be designed so that a person can effectively oversee it — and stop it.
What Article 14 requires
Providers must build in oversight measures so that natural persons can understand the system's capacities and limits, remain aware of automation bias, correctly interpret output, decide not to use it or to override it, and intervene or halt the system. For certain biometric systems, at least two people must confirm an action.
Who it binds
Providers design the oversight measures; deployers (Article 26) operationalise them by assigning competent, trained people with the authority to act.
Key points
- Oversight must be effective, not nominal — the person must be able to intervene and stop.
- Designed by the provider, operationalised by the deployer.
- Certain biometric identifications require two-person confirmation.
FAQ
What is human oversight under the EU AI Act?
It's the requirement that high-risk AI is designed so a competent person can monitor it, interpret its output correctly, and intervene or stop it — the guard against blindly trusting automation.
Related articles
Run the free 2-minute checker — deterministic, no sign-up.