EU nations have reached a provisional agreement on landmark legislation governing the use of artificial intelligence (AI), marking a significant step towards establishing a global standard for this rapidly evolving technology. The deal, finalized on December 8th, 2023, will address concerns around ethical issues, safety, and transparency in AI applications.

Here are some key points of the agreement:

• Prohibition of certain practices: The use of AI for facial recognition in public spaces will be banned unless authorized by law and under strict safeguards. This includes untargeted scraping of facial images from the internet or CCTV footage. Additionally, the legislation forbids social scoring systems and biometric categorization based on sensitive personal data like race, religion, and sexual orientation.
• Transparency and Explainability: Developers of high-risk AI systems, such as those used in critical infrastructure or healthcare, will be required to provide clear explanations of how their algorithms work. This is to ensure accountability and allow for independent assessment of potential biases or risks.
• Consumer protection: Consumers will have the right to be informed when they are interacting with AI systems and to object to automated decision-making based on their personal data. They will also have the right to lodge complaints and receive meaningful explanations if they believe their rights have been violated.
• Risk categorization: AI systems are categorized into four risk levels, with the highest-risk categories subject to the most stringent regulations. These categories include systems used in critical infrastructure, law enforcement, and healthcare.
• Fines and penalties: Companies found to be in violation of the new rules could face substantial fines, ranging from 7.5 million euros (approximately $8.1 million) or 1.5% of their global turnover to 35 million euros (approximately $38.5 million) or 7% of their global turnover.

The agreement has been welcomed by many as a necessary step in regulating AI development and deployment. However, some critics have expressed concerns that the regulations could stifle innovation and limit the positive potential of AI. Others, like Ella Jakubowska, senior policy advisor at EDRi, a European digital rights organization, criticize the legalization of live public facial recognition in specific circumstances.

The agreement is still provisional and requires formal approval from the European Parliament and Council. Nevertheless, it represents a significant milestone in AI governance and is likely to have a major impact on the global technology landscape.