In a historic move, the European Parliament has approved the groundbreaking Artificial Intelligence (AI) Act, aimed at safeguarding fundamental rights, democracy, and environmental sustainability while fostering innovation. The regulation, finalised in negotiations with member states in December 2023, received overwhelming support from members of the European Parliament (MEPs), with 523 votes in favour, 46 against, and 49 abstentions. The key objectives of the legislation are to shield fundamental rights and democracy from the risks posed by high-risk AI applications, enhance innovation, and position Europe as a frontrunner in the AI domain. The Act delineates obligations for AI systems based on their potential risks and impact levels.
Prohibited Applications and Law Enforcement Provisions
The Act prohibits several AI applications deemed threatening to citizens' rights, including biometric categorization systems based on sensitive characteristics and untargeted scraping of facial images for facial recognition databases. It also bans emotion recognition in workplaces and schools, social scoring, predictive policing solely based on profiling, and AI that manipulates human behaviour or exploits vulnerabilities. Law enforcement's use of biometric identification systems is restricted, except in narrowly defined situations with stringent safeguards. Real-time deployment requires specific authorization, such as targeted searches for missing persons or preventing terrorist activities. Post-facto use, known as "post-remote biometric identification," necessitates judicial authorization linked to criminal offences.
Obligations for High-Risk Systems and Transparency Requirements
High-risk AI systems, posing significant potential harm to health, safety, fundamental rights, environment, and democracy, are subject to clear obligations. These include risk assessment, transparency, maintaining use logs, accuracy assurance, and ensuring human oversight. Citizens are granted the right to lodge complaints about AI systems affecting their rights. General-purpose AI systems must meet transparency requirements, including compliance with EU copyright law and publishing detailed summaries of training data. Stricter measures apply to powerful AI models that could pose systemic risks, mandating model evaluations, risk assessments, incident reporting, and labelling of artificial or manipulated content.
Support for Innovation and SMEs
To foster innovation, regulatory sandboxes and real-world testing opportunities will be established at the national level, facilitating Small and medium-sized enterprises’ (SME) and startups' development and training of innovative AI solutions.
Voices from the Parliament
Brando Benifei, co-rapporteur from the Internal Market Committee, emphasised the Act's role in banning unacceptable AI practices, protecting workers' and citizens' rights, and promoting transparency. Dragos Tudorache, co-rapporteur from the Civil Liberties Committee, highlighted the Act's alignment with fundamental values and its implications for governance in the age of AI.
Next Steps and Background
The Act awaits final linguistic scrutiny and formal endorsement by the Council before its enactment. It will become fully applicable 24 months after publication, with specific provisions coming into force at varying intervals. The legislation responds directly to citizens' proposals from the Conference on the Future of Europe, emphasising competitiveness, safety, innovation, and responsible AI use.
The approval of the AI Act signifies a significant milestone in Europe's approach to regulating AI, striking a balance between innovation and safeguarding fundamental rights in the digital era.
Source: European Parliament
Image Credit: iStock
