Skip to content
EN

Which AI is prohibited under Article 5 of the AI Act?

The EU AI Act explainedLast updated 3 min read

Short answer

Since 2 February 2025 it has been prohibited to use AI to, among other things, read the emotions of employees or students, score people for their behaviour (social scoring), manipulate or exploit vulnerable people, categorise people by race, political opinion or sexual orientation based on biometrics, and build facial recognition databases by scraping images from the internet. From 2 December 2026, AI for non-consensual sexual deepfakes and child sexual abuse material is also banned. The maximum fine is €35 million or 7% of turnover.

The list of prohibited AI

Prohibited Example from an ordinary business
Harmful manipulation using hidden techniques An app that pushes customers into purchases they can't fully grasp
Exploiting vulnerability (age, disability, finances) Loans marketed aggressively to people in financial trouble
Social scoring Points for behaviour that lead to worse treatment elsewhere
Predicting crime based on profiling alone –
Facial recognition databases from scraped images Tools that find people from a photo online
Emotion recognition in the workplace and in education A camera that measures staff mood; emotion analysis in job interviews
Biometric categorisation by race, politics, beliefs or sexual orientation –
Real-time facial recognition in public spaces for law enforcement Police only, and with exceptions
Non-consensual sexual deepfakes and child sexual abuse material (from 2 December 2026) Generative AI with no safeguards

The one that affects most businesses

The ban on emotion recognition in the workplace and in education is the one ordinary businesses run into most often. It covers tools for video interviews, call centres and meetings that infer whether someone is happy, stressed, honest or engaged from their face or voice.

In GoalCube, prohibited practices are the first question when you register an AI use. If you answer yes, the use is stopped and the AI gets no data.

GoalCube AI MapFictitious data
What will the AI be used for?
Minimal risk

Approved automatically under your rules. Staff need sufficient AI literacy (Art. 4).

Limited risk

Users must be able to see they are talking to an AI (Art. 50). GoalCube has ready-made wording you can use.

High risk

Recruitment is listed in Annex III. A person confirms the use before the AI gets any data, and decides each case (Art. 14 and 26; Art. 22 GDPR).

Prohibited

Emotion recognition in the workplace is prohibited (Art. 5). GoalCube won’t help with it.

The example is simplified. In GoalCube you answer a few questions, and the classification follows fixed rules.

What to do

  1. Go through your AI tools.Especially tools for HR, monitoring, video and marketing.
  2. Ask the provider.Does the system read emotions, mood or personality from faces, voices or text?
  3. Stop anything prohibited.Switch the feature off or stop using it. There is no transition period.
  4. Write it down.Note in your AI register that you have checked for prohibited use.

What the law says

Article 5 AI Act – prohibited AI practices
Applies from 2 February 2025. New points (ba) and (bb) from 2 December 2026 under Regulation 2026/1744.
Article 99(3) AI Act – fines
Up to €35 million or 7% of worldwide annual turnover, whichever is higher. For small businesses, the lower of the two applies.
Commission guidelines on prohibited AI practices
Explains each ban with examples.

Questions and answers

Is it banned to measure employee wellbeing with AI?

An anonymous survey is not banned. But AI that infers emotions from faces, voices or body language in the workplace is banned – except for medical or safety reasons.

What about a fatigue alarm for drivers?

AI that detects tiredness for safety reasons can be exempt. If it is used to assess employees, that is a different matter.

Do the bans also apply to free tools?

Yes. The bans apply to the use, whether the tool is free or paid.

Sources

Written by The GoalCube editorial team. Last updated 11 October 2026. GoalCube helps you follow the EU AI Act and GDPR. It is not legal advice.

Be among the first.

We expect to launch in November 2026. Join the waitlist and we’ll let you know when you can get started.

Read next