After extensive negotiations between member states and the European Parliament, a preliminary deal has been reached on the world’s first Artificial Intelligence Act, designed to regulate the use of AI in an ethics-based manner. The agreement was struck after more than 35 hours of intense talks and will have far-reaching implications for regulatory efforts globally.
The focus of the negotiations was on tackling complex and technical issues related to AI, particularly in the use of real-time biometrics, including facial recognition, in public spaces. There was a significant debate over whether state authorities should be allowed to deploy AI-powered biometric systems, with MEPs arguing that such practices were intrusive and discriminatory, leading to a ban on certain applications of AI across EU territory. However, there was also strong debate over exemptions for law enforcement use, particularly in tracking down criminals and addressing threats to national security.
The compromise that emerged from these talks will require further fine-tuning in the coming days before being sent to the European Parliament for a new vote and then to the countries in the Council for approval. The law is expected to undergo final votes in early 2024 and will have a gradual period before becoming fully applicable.
The AI Act is a ground-breaking attempt to regulate the transformative technology in a human-centric, ethically responsible manner. It proposes a pyramid-like structure that splits AI-powered products into different categories based on the potential risk they pose to the safety and fundamental rights of citizens. AI systems with unacceptable risk for society will be banned across all EU territory, and companies that violate the rules will face significant fines.
The landmark legislation faced challenges throughout its development, particularly in addressing concerns over foundation models that power chatbots. There was skepticism from member states over these provisions, but a preliminary deal has now been reached, which introduces transparency requirements for chatbots to ensure users are aware they are interacting with an AI-powered system.
Overall, the AI Act represents a significant milestone in creating a regulatory framework for the development of AI that can be trusted, and once fully approved, will have profound implications for the use of AI across the EU.
Source
Photo credit www.euronews.com
