The AI Act Has Been Approved

Created in March 16, 2024 by: Garo Karh Bet

2024   ·   AI   ArtificialIntelligence   AiAct   EU   EuParliament   ·   AI

With overwhelming majority: The European Parliament approves the AI Act and here’s what you need to know.


The AI Act at a Glance

The AI Act is the first-ever legal framework on AI, which addresses the risks of AI and positions Europe to play a leading role globally.

It states that AI Systems of various applications are subject to analysis and classification based on the level of risk they pose to the users. This classification reflects in return the level of regulation these systems are going to underlie. Even though the existing legislation provides protection to some extent, it is still incapable of facing the numerous specific challenges these systems might carry.

“While many AI systems pose minimal risk, they need to be assessed.” 🤖

The way it works

The Regulatory Framework categorizes AI Systems into four main Risk Groups:

  1. Unacceptable Risk: AI that is considered a clear threat to the safety, livelihoods and rights of people.
  2. High Risk: AI used in critical infrastructures, education, etc.
  3. Limited Risk: Chatbots and GenAI.
  4. Minimal or No Risk: AI-enabled video games or spam filters.

All systems labeled as an unacceptable risk are considered a threat to people and will be prohibited, with limited exceptions for law enforcement purposes.

While applications posing high levels of risk will have to deal with strict obligations prior to their market introduction, minimal or no-risk AI systems are allowed to be used freely, with the vast majority of those currently operational within the EU falling into this category. On the other hand, AI systems in the third group have to comply with the required levels of transparency in order to promote trust. This includes informing users while interacting with a machine (i.e. chatbots), or labeling AI generated content accordingly.

Entry to force

The regulations governing general-purpose AI will take effect in May 2025, while those for high-risk systems, three years thereafter. They will be supervised by national authorities, backed up by the AI office inside the European Commission.

Sources: