The European Artificial Intelligence Law has been passed and will enter into force in two years – IT Pro – News

The AI ​​law will actually take effect in just over two years. The European Council gave its final approval this week. The regulation should clarify what is and is not permitted with artificial intelligence.

Now that the law of artificial intelligence Approved by the member states of the European UnionThe law will go into effect next month. After that, AI makers have two years to comply with this regulation. What AI makers should do depends on the risks of the model.

For example, AI that the EU sees as posing little risk should be transparent about the model used. While “high-risk AI systems” should only be deployed if logs are used, they are accurate and the risks are limited. Human supervision is also mandatory. This concerns AI systems used, for example, in critical infrastructure, education, work or basic social services.

Artificial intelligence systems that influence people’s thoughts and behavior are also prohibited. This also applies to social scoring systems that rely on artificial intelligence. Police are also not allowed to use AI to do predictive policing, and AI systems are not allowed to use biometric data to put people into specific categories such as race, religion, or orientation.

To ensure organizations comply with regulations, there will be an Artificial Intelligence Office and an Artificial Intelligence Council to assist the European Commission. In addition, there will be a panel of independent experts and an advisory forum where stakeholders can provide technical expertise to the AI ​​Governing Council and the Commission. Companies that do not comply with regulations risk fines. The amount is determined based on that company’s global annual sales volume.