Human Oversight (Article 14 EU AI Act)
Human oversight for high-risk AI
TL;DR
Human Oversight under Art. 14 EU AI Act is the obligation to design high-risk AI systems so that natural persons can oversee them effectively. Provider obligation (design): stop button, control mechanisms, clear outputs. Deployer obligation: people with competence + training + authority (Art. 26(2)).
What is Human Oversight (Article 14 EU AI Act)?
Providers must build in human oversight mechanisms (Art. 14(4)):
- Understanding of capabilities and limitations
- Awareness of automation bias
- Corrective intervention possible
- Stop button / override
- Outputs interpretable
Deployers must assign persons with competence, training and authority (Art. 26(2)). The reversal of the burden of proof under § 22 of the German General Equal Treatment Act (AGG) in conjunction with Art. 14, 26 AI Act: missing human oversight gives rise to AGG liability — even before Annex III applies.
Practical example
Practical requirements for an HR recruiting tool:
- HR staff trained in algorithmic decision-making
- Override option for every automated rejection
- Bias indicator display
- Audit trail of all decisions
Frequently asked questions
Is a 'stop button' enough?
No, that is the minimum baseline. Human oversight requires UNDERSTANDING + the ABILITY to intervene + avoidance of automation bias.
What is automation bias?
The tendency to follow automated recommendations uncritically. A person relies on the algorithm without questioning it — Art. 14(4)(b) requires awareness of this.
Do all employees have to understand HR AI?
No, only those with an oversight role. In practice: a dedicated 'AI oversight role' within the HR team.
See also
Practitioner note: This article is practice-oriented compliance documentation, not legal advice. We are compliance specialists, not a law firm. For legally binding advice, consult an admitted lawyer.