EU Reaches Landmark Deal on World’s First Comprehensive AI Act

European Union negotiators have finalized a provisional agreement on the Artificial Intelligence Act. The deal was struck after marathon talks concluded in Brussels. This legislation is the first of its kind globally.

The act aims to regulate artificial intelligence based on its potential risks. Officials from the European Commission, Parliament, and member states approved the text. According to Reuters, the agreement establishes a clear legal framework for AI development and use.

Key Provisions and Risk-Based Bans

The new law categorizes AI systems by the level of threat they pose. Applications deemed “unacceptable risk” will be banned. This includes AI for social scoring by governments.

It also bans cognitive behavioral manipulation and predictive policing based on profiling. Certain uses of remote biometric identification in public spaces by law enforcement are restricted. Exceptions exist for serious crimes like kidnappings or terrorist threats.

High-risk AI, such as that used in critical infrastructure, must meet strict obligations. These include rigorous risk assessments and high-quality data sets. Citizens will have a right to launch complaints about AI systems.

Governance and Penalties for Violations

A new European AI Office will be created within the Commission. This body will oversee the implementation of the rules for general-purpose AI models. It will enforce compliance across the single market.

Violations of the act could lead to significant financial penalties. Fines can range from 7.5 million euros or 1.5% of turnover to 35 million euros or 7% of global turnover. The severity of the fine depends on the infringement and the company’s size.

The rules for foundational models, like GPT-4, will be tiered. All models must meet basic transparency requirements. Those posing “systemic risks” will face more stringent evaluations.

Broader Impact on the Tech Industry

The EU’s law is expected to have a global impact, similar to the GDPR data law. Tech companies operating in the EU’s vast market must comply. This could set a de facto standard for other regions developing their own rules.

Industry groups have expressed both support and concern. Some worry the regulations could stifle innovation in Europe. Proponents argue it creates necessary trust and legal certainty for both businesses and citizens.

The final legal text is now subject to formal approval by the European Parliament and Council. The Act is expected to come into force in 2026, following a phased implementation period. This gives companies time to adapt to the new requirements.

The EU AI Act represents a historic step in governing transformative technology. Its risk-based approach aims to balance innovation with fundamental rights. The world will be watching as this framework shapes the future of artificial intelligence.

Thought you’d like to know

What is the EU AI Act?

The EU AI Act is a comprehensive law to regulate artificial intelligence. It classifies AI systems by risk level and sets rules for their use. It is the first major attempt to create a legal framework for AI.

What AI applications are banned?

The Act bans AI considered an unacceptable risk. This includes government social scoring and manipulative AI. It also heavily restricts real-time remote biometric identification in public.

When will the AI Act become law?

The provisional deal needs final approval from EU lawmakers. This process is expected to be completed in 2024. The law will then be phased in over two years.

How will the Act affect companies outside the EU?

Any company offering AI products or services in the EU market must comply. This gives the regulation a global reach, similar to the EU’s GDPR privacy law.

What are the penalties for breaking the rules?

Fines can be very large, up to 35 million euros or 7% of a firm’s global annual turnover. The exact amount depends on the type of violation and the company’s size.

Trusted Sources

Reuters, Associated Press, European Commission, BBC News.