The EU's New AI Legislation – Everything You Need to Know

The EU's new legislation, the first of its kind in the world, is extensive, with four different safety levels to comply with. All to safeguard the security of member states. We've gathered and compiled the available information for you.

Share
The EU's New AI Legislation – Everything You Need to Know

In a historic vote, the European Parliament has voted to introduce legislation on artificial intelligence (AI), marking one of the first comprehensive regulatory frameworks for AI in the world. With an overwhelming majority of 523 votes in favor, 46 against, and 49 abstentions, Parliament has taken a step toward regulating the technology, aiming, among other things, to prevent mass surveillance and protect citizens' personal privacy.

Why Do We Need Legislation?

At a time when artificial intelligence (AI) plays an increasingly large role in our lives, the EU has taken important steps to ensure the technology is used in a way Europeans can trust. AI technology offers enormous opportunities but also carries risks that must be managed to protect citizens and society.

Challenges and Risks of AI

Although many AI systems are intended to have a positive impact, there are cases where the use of AI can lead to unwanted consequences. One example is the difficulty of understanding how an AI system arrived at a particular decision, which can make it challenging to assess whether the decision is fair. This opacity can lead to situations where individuals are disadvantaged, for example, in recruitment processes or when applying for public benefits.

The Purpose and Goals of the AI Act

The AI Act is the EU's response to these challenges and aims to create a safe and fair environment for the use of AI, by:

- Managing risks specifically created by AI applications,
- Prohibiting AI methods that pose unacceptable risks,
- Establishing a list of high-risk applications,
- Setting clear requirements for AI systems used in high-risk applications,
- Establishing specific obligations for deployers and providers of high-risk AI applications,
- Requiring a conformity assessment before a given AI system is put into use or placed on the market,
- Introducing enforcement measures after a given AI system has been placed on the market,
- Establishing a governance structure at the European and national level.

Source - The European Commission

Categorization of Risks and Regulation

Source: The European Commission

The new legislation divides AI systems into four risk categories:

Unacceptable risk applies when AI "is considered to pose a threat to people's safety, livelihoods, and rights" and will be prohibited. For example, AI systems for social scoring, which exist in China today.

High risk includes AI systems used in critical infrastructure, education, employment, or in connection with migration, asylum, and border control. These AI systems will be subject to "strict obligations" before being placed on the market. Law enforcement is assessed to fall into this risk level.

Limited risk covers, for example, chatbots, where one requirement is that users must be aware that they are interacting with a machine.

Minimal risk allows for free use, as with AI-powered video games or spam filters.

The Transparency Requirement

For AI systems not considered high-risk, there is a requirement that content created or modified by these systems be clearly labeled as AI-generated. This covers a broad range of content types, including text, images, audio, and video files, to inform users of how the information was created.

For more advanced and potentially higher-risk AI systems, such as GPT-4, stricter regulations are introduced in addition to the transparency requirements. These systems must undergo extensive reviews to identify and address any risks they may pose. Any suspicious behavior or output generated by these powerful AI models must be reported directly to the European Commission for further evaluation.

With this legislation, the EU hopes not only to protect its citizens but also to inspire global standardization of AI regulation. By collaborating with other countries working on similar legislation, the EU aims to create a unified legal framework that promotes the development of AI technology in a safe and ethical way. The AI Act is expected to come into force gradually, with full implementation planned for 2025.

Sources

The European Parliament - Link

The European Commission - Link

The European AI Office - Link