EU Council: First Global AI Regulation

yippy
2 min read
May 22, 2024 8:02:42 AM

The Council of the European Union has officially approved the pioneering Artificial Intelligence (AI) Act, setting the first-ever global standard for AI regulation.This groundbreaking legislation adopts a 'risk-based' approach, imposing stricter rules on AI systems that pose higher risks to society.

The AI Act aims to foster the development and adoption of safe and trustworthy AI systems throughout the EU's single market by both private and public sectors. It seeks to ensure the protection of fundamental rights for EU citizens while also encouraging investment and innovation in AI technologies. The Act applies solely to areas governed by EU law and includes exemptions for systems used exclusively for military, defense, and research purposes.

"This landmark law addresses a global technological challenge while creating opportunities for our societies and economies. With the AI Act, Europe emphasizes the importance of trust, transparency, and accountability in dealing with new technologies," said Mathieu Michel, Belgian Secretary of State for Digitisation.

High-Risk and Prohibited AI Practices

The AI Act classifies AI systems based on their risk levels. Systems presenting limited risk face minimal transparency obligations, while high-risk AI systems must meet stringent requirements to access the EU market. AI applications like cognitive behavioral manipulation and social scoring are banned due to their unacceptable risk levels. Additionally, the Act prohibits AI use in predictive policing based on profiling and systems that categorize individuals by race, religion, or sexual orientation using biometric data.

General Purpose AI Models

The legislation also addresses general-purpose AI (GPAI) models. Those posing no systemic risks will have limited requirements, primarily related to transparency. However, GPAI models with systemic risks must comply with more rigorous standards.

New Governance Structure

To ensure effective enforcement, the AI Act establishes several governing bodies:

  • An AI Office within the Commission to enforce rules across the EU
  • A scientific panel of independent experts to support enforcement activities
  • An AI Board comprising member state representatives to advise the Commission and member states on consistent application of the Act
  • An advisory forum for stakeholders to provide technical expertise to the AI Board and the Commission

Penalties and Compliance

The Act stipulates fines for non-compliance, calculated as a percentage of the offending company's global annual turnover or a predetermined amount, whichever is higher. SMEs and start-ups will face proportional administrative fines.

Transparency and Fundamental Rights Protection

High-risk AI systems deployed in public services must undergo a fundamental rights impact assessment before use. The regulation mandates increased transparency regarding the development and use of high-risk AI systems. Public entities using high-risk AI systems and emotion recognition systems must register these in the EU database and inform individuals when they are being exposed to such systems.

Supporting Innovation

The AI Act promotes an innovation-friendly legal framework, encouraging evidence-based regulatory learning. It includes provisions for AI regulatory sandboxes, allowing the testing and validation of innovative AI systems in controlled, real-world conditions.

Next Steps

The legislative act will be signed by the presidents of the European Parliament and the Council and published in the EU's Official Journal in the coming days. It will enter into force 20 days after publication, with its provisions becoming applicable two years later, with certain exceptions.

Background

The AI Act is a critical component of the EU's strategy to promote safe, lawful AI that respects fundamental rights. The proposal was submitted by Commissioner Thierry Breton in April 2021, with Brando Benifei and Dragoş Tudorache serving as European Parliament rapporteurs. A provisional agreement was reached on 8 December 2023.

Topics: AI law