(
European
Union
Artificial
Intelligence Act) An EU regulation that established a legal framework for AI within the EU. Enacted in 2024, it includes AI used in several sectors but not the military. Following is a brief summary of the risk categories. See
AI legislation.
Unacceptable
Applications that manipulate human behavior, including social scoring, are banned.
High Risk
Applications that pose risk to health and safety must be further evaluated.
General Purpose
Including foundation models such as ChatGPT, the internal architecture must be made available under open source so everyone can understand it.
Limited Risk
Generative AI systems must inform users that AI is being used to create text, images and videos.
Minimal Risk
AI-enhanced applications such as video games, spam filters and everyday processing have little requirements.