Legal framework
AI Act
In august 2024 a new law on AI usage was introduced: the AI Act. The goal of the law is to guarantee justified and ethical use of artificial intelligence while also stimulating innovation and competition at the same time. In short, the law consists of two parts: taking inventory of the risks of AI and making its users literate.
Risks
You are expected to take inventory of the AI tools you use and classify them.
Image: European Commission
The law classifies AI into four levels of risk:
- Unacceptable risk
Tools that fall in this category are illegal to use. If these are already in use, they must be disabled.
Think of:- Biometric systems for identification at a distance in public spaces.
- AI-emotion recognition at work and school.
- Systems that can mislead or manipulate people.
High risk
Think of the following in education:- Systems which are used to assess which students are admitted
- Adaptive learning materials
- Systems to detect fraud
Limited risk
For example: chatbots or AI systems that can generate text, images, audio or video. You are expected to be transparent about the use of these systems.Minimal risk
For example: spam filters or video games.