Commission Guidelines on Prohibited AI Practices in Education
The European Commission has published Guidelines on Prohibited Artificial Intelligence (AI) Practices as stipulated in the European Artificial Intelligence Act (AI Act), which entered into force in August 2024. These guidelines provide legal explanations and practical examples to help stakeholders understand the requirements of the regulation, which classifies AI systems into four levels of risk: minimal to no risk, transparency risk, high risk and unacceptable risk.
Among the prohibited practices it highlights the recognition of emotions in educational and work environments, except in cases of medical or security use. The European Commission considers that this technology can be intrusive and discriminatory, affecting the privacy, human dignity and freedom of thought of teachers and students.
What is prohibited in education?
Article 5 of the AI Act lists a number of practices that are considered an unacceptable risk and are therefore banned in the EU:
🚫 Harmful manipulation, and deception: AI systems that use subliminal or manipulative techniques that distort people's behaviour and may cause significant harm are prohibited.
🚫 Harmful exploitation of vulnerabilities: It is prohibited to use AI to exploit a person's age, disability or social and economic situation in order to influence his or her behaviour in a damaging way.
🚫 Social scoring: It is not permitted to classify persons or groups according to their behaviour or personal characteristics where this would result in unfair or disproportionate treatment.
🚫 Individual criminal offence risk assessment and prediction: It prohibits systems that predict the likelihood of a person committing a crime on the basis of personal profiles or characteristics alone, unless they complement an objective human assessment.
🚫 Untargeted scraping to develop facial recognition databases: The development of facial recognition databases through the mass collection of images from the Internet or surveillance cameras is banned.
🚫 Emotion recognition: In educational and work environments, the use of AI to infer emotions is prohibited, except for medical or safety reasons.
🚫 Biometric categorisation: It is not permitted to use AI to classify individuals on the basis of race, political or religious beliefs, sexual orientation or other biometric data, except in specific cases of data tagging.
🚫 Real-time remote biometric identification: The use of remote biometric identification in public spaces for law enforcement purposes is prohibited, except in cases of victim tracing, threat prevention or criminal investigations under strict authorisation.
The regulation places special emphasis on protecting the educational sphere, prohibiting the use of AI for the purpose of:
❌ Assessing students' attention and motivation through the recognition of emotions.
❌ Use AI in admission or certification assessments that detects emotions such as emotional arousal or anxiety.
❌ Implementing emotion recognition systems for teachers in the work environment.
Exemptions and permitted uses
AI-based eye-tracking systems to detect the use of unauthorised material in online examinations, provided they do not infer emotions.
Role-playing games and training simulations in which the AI detects emotions without the results affecting the student's certification.
Therapeutic use of AI in educational environments to address medical or security issues.
