Last December, the European Parliament reached a provisional agreement with the Council on the text of the new Artificial Intelligence (AI) Act, which will be pioneering worldwide.
The Parliament's priority is to ensure that AI systems used in the EU are secure, transparent, traceable, non-discriminatory and environmentally friendly. Furthermore, AI systems should be supervised by actual people and not governed by automation.
The new rules establish obligations for providers and users depending on the risk level of the AI used. Three levels of risk have been identified: unacceptable, high and limited.
In the first group are AI systems considered a threat to people. They include those aimed at the cognitive-behavioral manipulation of specific vulnerable people or groups (like children); those who develop a "social score" to classify people according to their behavior, socioeconomic status or personal characteristics; biometric identification and categorization of people and remote and real-time biometric identification systems, such as facial recognition. These AIs will be banned.
AI systems that negatively affect security or fundamental rights are categorized as high risk. Here we find AI that is used in products subject to EU safety legislation (toys, aviation, cars, medical devices, etc.) as well as generative AI, such as the famous ChatGPT application. The latter will have to comply with transparency requirements, that is, reveal that the content is generated by AI and publish summaries of the copyrighted data used in the training of the AI. It must also actively avoid the generation of illegal content.
Finally, limited risk AI systems will need to meet minimum transparency requirements that allow users to make informed decisions. For instance, users must be aware when interacting with an AI, including systems that generate or manipulate image, audio or video content, for example, deepfakes.
The agreed text will now have to be formally adopted by both Parliament and the Council to become EU law. This law hopes to become a pioneering and influential framework worldwide for a field so promising, but yet So challenging to both privacy and intellectual property rights.