An unprecedented law to regulate Artificial Intelligence in the European Union has come into force
The AI Act is the first legal framework on Artificial Intelligence, which addresses the risks of AI and positions Europe to play a leading role globally.
This is the first comprehensive legal framework on AI worldwide. The aim of the new rules is to promote trustworthy AI in Europe and beyond, ensuring that AI systems respect fundamental rights, safety and ethical principles and addressing the risks of very powerful and impactful AI models.
The new rules:
- Address the risks specifically created by AI applications
- Prohibit AI practices that pose unacceptable risks
- Determine a list of high-risk applications
- Define clear requirements for AI systems for high-risk applications
- Define specific obligations for high-risk AI application providers and deployers
- Require a conformity assessment before a particular AI system is put into service or placed on the market
- Enforce law enforcement after a particular AI system is brought to market
- Establish a governance structure at European and national level
The AI Act divides AI systems into a hierarchy of risks: minimal risk, specific transparency risk, high risk, and unacceptable risk.
- Minimal risk: there are systems that make recommendations – for example, advertisements based on the content consumed – and do not pose any problem for citizens.
- Specific risk: Enter chatbots, which must clearly inform users that they are interacting with a machine and if they are deep fakes they must be identified that way.
- High risk: Enter algorithms that truly mess with people's lives and can cause obstacles to the development or lives of citizens, including those that recruit, assess whether someone is entitled to a loan or manage autonomous robots. There they will have concrete obligations of human control, login and activity registration, among others.
- Unacceptable risk: Includes toys for minors that use voice assistance to encourage dangerous behavior, systems of governments and organizations that do "social scoring," and predictive policing applications.
The 27 EU countries have until August 2nd, 2025 to designate national competent authorities to supervise the application of these rules and carry out market surveillance.
Watch the video: https://twitter.com/vonderleyen/status/1818907654921535707
Know more:
AI Act: Regulamento europeu da Inteligência Artificial entra em vigor (sapo.pt)
