The European Parliament and the Council have reached a preliminary agreement on the Artificial Intelligence Act, a crucial step towards ensuring that AI in Europe meets safety and ethical standards while fostering business growth. The agreement aims to protect fundamental rights, democracy and environmental sustainability from high-risk AI threats.
The law bans AI applications that pose risks to citizens’ rights, such as certain biometric categorisation systems and AI that manipulates behaviour or exploits vulnerabilities. It also sets strict conditions for the use of AI in law enforcement, including judicial authorisation for biometric identification in public spaces.
High-risk AI systems, including those affecting health, safety and fundamental rights, require rigorous impact assessments and transparency. This includes AI influencing elections or voter behaviour.
The law encourages innovation and supports SMEs through regulatory sandboxes and real-world testing. Non-compliance will result in significant fines, depending on the offence and the size of the company.
This provisional agreement, which has been well received by key stakeholders, emphasises a human-centred approach to AI, protects SMEs and ensures robust rules for AI models. Formal adoption of this legislation by the Parliament and Council will position the EU at the forefront of ethical AI development.