Strict New Rules for High-Risk AI Applications

The European Union has formally approved its landmark Artificial Intelligence Act. This sweeping new law creates the world’s first comprehensive legal framework for AI. It was finalized this week after months of intense negotiations. The act categorizes AI systems by risk and bans applications deemed unacceptable. Its passage signals a major shift toward global AI governance.

According to Reuters, the act sets a precedent other nations are already examining. Officials aim to balance innovation with fundamental rights protection. The rules are expected to fully apply within two years.

The core of the EU AI Act is a risk-based pyramid. At the top are prohibited applications. These include AI for social scoring and real-time biometric surveillance in public. The use of emotion recognition in workplaces and schools is also banned.

High-risk AI, like systems used in hiring or critical infrastructure, faces strict obligations. Developers must ensure data quality and provide detailed documentation. Human oversight and robust risk assessments are mandatory. Citizens will have a right to explanations for AI-driven decisions.

Violations carry severe penalties. Fines can reach up to 7% of a company’s global turnover. This creates a powerful incentive for compliance across industries.

Global Race to Shape the Future of AI Governance

The EU’s move accelerates a global regulatory race. The United States issued an executive order on AI safety last year. Japan and other nations are crafting their own guidelines. However, the EU’s approach is notably more centralized and stringent.

This creates a potential “Brussels effect.” Much like the GDPR did for data privacy, the AI Act could become a de facto global standard. Multinational companies may adopt its requirements worldwide for simplicity. This gives the EU significant influence over the technology’s future development.

Analysts note a key challenge will be enforcement. National authorities will need significant resources. The pace of technological change may also outstrip the law’s provisions.

The finalization of the EU AI Act marks a pivotal moment in tech governance, setting a concrete benchmark for responsible AI development that the rest of the world must now address.

Thought you’d like to know

What are the main categories of AI under the new EU law?

The law uses four risk tiers. It bans unacceptable-risk AI outright. It strictly regulates high-risk AI used in critical areas. Limited-risk AI has transparency duties, like chatbots disclosing they are AI. Minimal-risk AI faces no new rules.

When will the AI Act fully take effect?

Prohibitions on unacceptable AI practices apply six months after the law enters force. Rules for general-purpose AI models start after one year. The full set of obligations for high-risk systems applies two years after enactment.

How does this affect companies outside the European Union?

The rules apply to any provider placing AI systems on the EU market or whose output is used within the EU. This means global tech firms must comply if they operate in the 27-nation bloc, impacting their worldwide product development.

What is considered ‘high-risk’ AI?

High-risk AI includes systems used in biometric identification, critical infrastructure, education, employment, essential services, law enforcement, and migration management. These systems require conformity assessments before deployment.

Does the law address generative AI like ChatGPT?

Yes. It introduces specific rules for general-purpose AI models. Stricter requirements are imposed for models with “systemic risk,” based on their computational power. These include mandatory risk evaluations and incident reporting.