.png "blank Coverage")
After intense negotiations, lawmakers in Brussels have reached a "provisional agreement" on the European Union's proposed Artificial Intelligence Act (AI Act). This legislation is poised to become the world's first comprehensive set of rules governing AI, potentially setting a precedent for other regions seeking to regulate AI technologies.
The provisional agreement outlines obligations for "high-impact" general-purpose AI (GPAI) systems that meet specific benchmarks, including risk assessments, adversarial testing, and incident reporting. Transparency is a key requirement, mandating the creation of technical documents and detailed summaries of the content used for training AI systems. Notably, citizens will have the right to file complaints about AI systems and receive explanations for decisions made by "high-risk" systems affecting their rights.
The framework includes provisions for fines in case of rule violations, with penalties varying based on the nature of the violation and the size of the company. Fines can range from 35 million euros or 7 percent of global revenue to 7.5 million euros or 1.5 percent of global revenue or turnover.
The AI Act bans the use of AI in specific applications, such as scraping facial images from CCTV footage, categorization based on "sensitive characteristics" like race or sexual orientation, emotion recognition at work or school, and the creation of "social scoring" systems. Additionally, it prohibits AI systems that manipulate human behavior to circumvent free will or exploit people's vulnerabilities. The legislation also includes safeguards and exemptions for law enforcement's use of biometric systems.
While a provisional agreement has been reached, further negotiations and votes by Parliament's Internal Market and Civil Liberties committees are required before the AI Act becomes law. A final deal is expected to be reached before the end of the year, and the legislation may not come into force until 2025 at the earliest. The EU's AI Act has undergone multiple revisions since its initial draft in 2021, adapting to the evolving landscape of AI technologies, such as generative AI tools like ChatGPT and Stable Diffusion. The negotiations have been particularly contentious, with debates over regulations for live biometric monitoring and general-purpose AI models contributing to delays in finalizing the agreement.
