By 
The European Union has released the final text of the Artificial Intelligence Act, detailing the key deadlines for compliance with the world’s first comprehensive AI regulation.
Following its approval by the EU Council in May, the legislation will come into effect on August 1, 2024. By August 2, 2026, AI developers operating within the EU must fully adhere to the AI Act provisions. Additionally, EU member states are required to have at least one AI regulatory sandbox operational at a national level by this date.
Notably, biometrics developers, providers, and users must be aware of an earlier deadline: February 2, 2025. By this date, AI applications deemed to pose “unacceptable risk” will be banned. Prohibited applications include biometric categorization based on sensitive characteristics, emotion recognition in workplaces and schools, untargeted scraping of facial images from the web or CCTV footage to create facial recognition databases, and real-time remote biometric identification in public spaces. Exceptions are made for law enforcement in specific circumstances.
Another crucial deadline is May 2, 2025, when providers will receive their codes of practice, outlining how they can demonstrate compliance. These codes will be issued by the newly-established AI Office, responsible for ensuring that classification rules and procedures remain current.
General-purpose AI systems, such as ChatGPT, must adhere to copyright and transparency rules by August 2025.
By August 2027, one year after the AI Act’s general applicability, developers of high-risk AI systems will need to comply with additional obligations outlined in Annex I of the AI Act. High-risk AI systems, including remote biometric identification systems, can be used in the EU market as long as they meet specific requirements.
Non-compliance with certain AI practices can result in substantial fines: up to 35 million euros (US$38.1 million) or seven percent of worldwide annual turnover, whichever is higher.
The European Commission has also provided a schedule for reviewing the regulation and amending the list of high-risk AI systems.
Comments