Earlier this month, the European Parliament took a significant step towards shaping global standards in AI regulation by approving its draft proposal for the AI Act, a comprehensive piece of legislation that has been two years in the making. Expected to be approved before the end of the year, this act will become the world’s first dedicated legislation to regulate AI across a variety of sectors. Rather than regulating AI itself, the legislation focuses on the specific domains of society where AI is used, categorizing them based on levels of risk: unacceptable, high, and limited.
The AI Act aims to completely prohibit systems that pose an “unacceptable risk” to fundamental rights or EU values, such as cognitive behavioral manipulation of people or specific vulnerable groups, social scoring, and real-time and remote bio metric identification systems such as facial recognition technology.
Systems classified as “high risk” will be subject to disclosure requirements and registration in a special database. This category includes: AI applications controlling access to critical services in education, employment, financing, healthcare, and other essential areas. These high-risk systems will undergo assessment before being marketed and throughout their lifecycle to ensure the safety and protection of fundamental rights.
High-risk AI systems have been placed in to two categories. The first includes AI systems used in products falling under the EU’s product safety legislation such as: toys, aircraft, cars, medical devices, and elevators. The second category focuses on eight specific areas requiring registration in an EU database.
These areas include:
Bio metric Identification
Critical Infrastructure Management
Education and Vocational Training
Employment and Worker Management
Access to Essential Services
Law Enforcement
Migration and Border Control
Legal Interpretation and Application of the Law
Limited-risk systems will have minimal transparency requirements, and operators of generative AI systems, like text or image-generating bots, will have to disclose that users are interacting with a machine.
The European Parliament’s approval of the AI Act draft proposal marks a significant milestone in global AI regulation, emphasizing human oversight of AI systems to prevent harmful outcomes rather than relying solely on automation. It aims to prohibit systems with unacceptable risks while subjecting high-risk systems to disclosure, registration, and assessment processes. Ultimately, the priority is to ensure that AI systems used within the EU are safe, transparent, traceable, non-discriminatory, and environmentally friendly. The resulting law is expected to be enforced by the end of 2023, followed by a two to three year window – after which full compliance will be expected.
Sources
https://theconversation.com/eu-approves-draft-law-to-regulate-ai-heres-how-it-will-work-205672
