Recently, the European Union reached a landmark preliminary agreement on the use of artificial intelligence (AI) after 15 hours of negotiations. This political agreement marks the first country in the world to enact AI regulation laws. The agreement covers regulations for the use of AI in biometric surveillance, as well as how AI systems, including ChatGPT, will be regulated.
This agreement requires underlying models (such as ChatGPT) and general artificial intelligence systems (GPAI) to comply with transparency obligations before going public. This includes the development of technical documentation, compliance with EU copyright law, and the publication of detailed summaries of the content used for training. High-impact underlying models with systemic risk must undergo model evaluation, assess and mitigate systemic risks, conduct adversarial testing, report major incidents to the European Commission, ensure cybersecurity, and report on their energy efficiency.
The agreement also stipulates that real-time biometric surveillance may only be used in public places, in cases of victims of specific crimes, to prevent real, current or foreseeable threats (such as terrorist attacks), and to search for persons suspected of the most serious crimes. The protocol prohibits cognitive-behavioral manipulation, undirected acquisition of facial images from the internet or CCTV footage, social scoring, and biometric classification systems used to infer political, religious, philosophical beliefs, sexual orientation, and ethnicity.
Consumers will have the right to file a complaint and receive a meaningful explanation, with fines ranging from 7.5 million euros or 1Ranging from 5% to 35 million euros or 7% of global turnover.
The legislation is expected to come into force early next year and will begin to apply in two years once formally approved by both parties. These ambitious European AI rules come at a time when companies like OpenAI continue to discover new uses for their technology, sparking praise and concern.