On Wednesday, the document was adopted in the European Parliament by an overwhelming majority (499 votes in favour, 28 against, and 93 abstentions). The AI Act – – is “the world’s first set of comprehensive rules for managing the risks of artificial intelligence.” This includes e.g. o possible supervision over people, issues of privacy, non-discrimination and the promotion of artificial intelligence applications that are in line with the values of the European Union.
EU rules for AI. Minority Report technology not in the EU
As we wrote on Tuesday, the act on artificial intelligence divides AI-based tools into several categories depending on the potential risk associated with their use. The use of AI will be banned for social evaluation (classifying citizens based on their behavior), emotion recognition and remote biometric surveillance both in real time and after the fact. This last point aroused a lot of emotions before the vote due to the controversial proposal of the amendment by the European People’s Party, which was eventually rejected by the parliament.
Pursuant to the AI Act, systems using biometric categorization based on sensitive features (e.g. race, gender, religion, origin) will also be banned. It will also not be allowed to train face recognition systems based on photos of random people downloaded from the network (or recorded). It should also be unacceptable use of AI systems for criminological prediction (based on profiling, location or previous criminal behavior). This practice brings to mind scenes from the famous “Minority Report”, where technology allowed the services to stop future criminals before they committed criminal acts. Even if artificial intelligence is able to predict crimes on its own, services in EU countries will not be able to use it.
The bot will inform you that it is not human
Included in the high-risk category are: AI systems that can be used to influence voters and election results. Recommendation systems, which are extremely popular today, are also included here, but only those used in social networking sites, used by at least 45 million people. They will be able to work, but under certain restrictions. The changes will also affect generative artificial intelligence. Chatbots, such as ChatGPT, are to inform users that the content they generate comes from AI, and image generation programs (eg Midjourney) will have to mark images as artificially created. Companies responsible for such tools will also have to ensure that illegal content is not generated. If, in turn, enterprises train AI on copyrighted data, they will have to inform about it in detailed summaries.
Consultations with the EU Council on the final shape of the new law are to begin on Wednesday.
Source: Gazeta
Mabel is a talented author and journalist with a passion for all things technology. As an experienced writer for the 247 News Agency, she has established a reputation for her in-depth reporting and expert analysis on the latest developments in the tech industry.
