EU passes comprehensive new set of rules to regulate AI

The EU has agreed on the details of the AI law. This far-reaching set of rules applies to all those who develop and use artificial intelligence and is intended to be a milestone. After months of debate on the regulation of companies such as OpenAI, lawmakers from the three branches of EU government spent a total of over 36 hours drafting the new law.

Most comprehensive set of rules of its kind

The EU AI law is the most comprehensive set of rules of its kind for this technology. Among other things, it prohibits biometric systems that identify people based on sensitive characteristics such as sexual orientation and race, as well as the indiscriminate tapping of faces from the internet. The legislative bodies also agreed that law enforcement authorities should be able to use biometric identification systems in public spaces for certain criminal offenses.

New transparency requirements

In addition, new transparency requirements have been included for all general purpose AI models, such as OpenAI's GPT-4, which powers ChatGPT, as well as stricter rules for "very powerful" models. “The AI Act sets rules for large, powerful AI models, ensuring they do not present systemic risks to the Union,” says DragosTudorache, Member of the European Parliament and one of the two co-rapporteurs who led the negotiations.

Companies that do not comply with the regulations can be fined up to 7 percent of their global turnover. The bans on prohibited AI will come into force in six months, the transparency requirements in 12 months and the full set of rules in around two years.

Measures to facilitate the protection of copyright holders from generative AI and the obligation for general AI systems to be more transparent about their energy consumption have also been included.

Source: www.wired.com