Glossary · EU AI Act: foundations
Prohibited AI practices (AI Act)
Also known as: prohibited AI, unacceptable risk, banned AI practices
Prohibited AI practices are the uses of AI that the EU AI Act bans outright because they pose an unacceptable risk to fundamental rights — among them manipulative techniques, exploitation of vulnerabilities, social scoring, untargeted scraping of facial images and emotion recognition at work and in education. The prohibitions apply since February 2, 2025.
- AI regulation
- EU
In one sentence
Prohibited AI practices are the uses banned by Art. 5 of the AI Act since February 2025, such as social scoring and emotion recognition at work.
Example
An employer may not use a webcam-based system that infers employees’ emotions during meetings, unless it is used for medical or safety reasons.
The prohibited practices
- Subliminal, manipulative or deceptive techniques that materially distort behavior and cause significant harm.
- Exploiting vulnerabilities due to age, disability or a social or economic situation.
- Social scoring leading to detrimental or disproportionate treatment.
- Predicting the risk that a person commits a crime based solely on profiling or personality traits.
- Building facial recognition databases by untargeted scraping of images from the internet or CCTV.
- Emotion recognition in the workplace and in education, except for medical or safety reasons.
- Biometric categorization to infer sensitive characteristics such as race, political opinions or sexual orientation.
- Real-time remote biometric identification in publicly accessible spaces for law enforcement, with narrow exceptions.
The Digital Omnibus on AI adds prohibitions on AI that generates non-consensual intimate imagery and child sexual abuse material, applying from December 2, 2026.
How it applies
- Technical documentation: Intended purpose and foreseeable misuse matter. Describe in the instructions what the system must not be used for, especially where a general capability — such as emotion recognition — would be prohibited in some settings.
Compared with the USA, Canada and China
The USA has no federal list of prohibited AI uses; some states prohibit specific uses — Texas, for example, prohibits developing or deploying AI with the intent to incite self-harm, harm to others or crime, or to discriminate unlawfully against a protected class. Canada has no equivalent. China does not prohibit practices by name in the same way, but its generative AI and deep synthesis rules prohibit generating content that endangers national security, spreads false information or infringes the rights of others.