Knowledge center

The first step towards the first law on artificial intelligence

On December 9 (Vietnam time), after 36 hours of negotiations, European Union (EU) member states agreed on a set of rules governing artificial intelligence (AI) models such as ChatGPT. .

------

These are the world's first comprehensive artificial intelligence (AI) rules, paving the way for regulatory oversight of the technology used in today's popular AI services.

Negotiators from the European Parliament and the bloc's 27 member states overcame major differences to sign a political agreement that is the basis for the "Artificial Intelligence Act".

"Agreement has been reached," European Commissioner Thierry Breton, posted on social network X. "Europe becomes the first continent to set clear rules for the use of AI."

This result was achieved after long closed negotiations over the past week, with the first session lasting 22 hours before the second round began on the morning of December 8.

The EU took an early lead in the global race to develop AI rules, publishing the first draft of the code in 2021. However, the recent boom in AI has made European officials nervous. Europe must be quicker in this matter.

The European Parliament will still need to vote early next year, however, with the newly reached agreement, it is just a formality, according to Italian lawmaker Brando Benifei.

Currently, the US, UK, China and global alliances such as the G7 have put forward proposals to regulate AI, although they have not yet caught up with Europe.

According to AFP, EC President Ursula von der Leyen welcomed the agreement, evaluating the EU's AI law as the only legal framework for developing trustworthy AI "for the safety as well as the fundamental rights of people." people and businesses".

The Guardian newspaper said that there is still little detailed information related to the AI law announced by the EU. According to The Verge website, this law - which the EU hopes can serve as a "standard" for other countries/regions around the world to consider adopting similar regulations - is expected as soon as possible. It will take until 2025 to take effect.

The "AI Act" was rushed into being by the EU after the ChatGPT chatbot exploded in late 2022. While ChatGPT's ability to generate essays and poems is an impressive display of rapid advances of AI, but experts are worried about how this technology can be abused. Along with that, derivative AI software, including Google's Bard chatbot, can quickly generate text, images and sounds from simple commands in everyday language.

According to regulations, Europe's AI law still needs to be passed by member states and parliament to take effect, but this agreement is considered the last major hurdle.

The EU is not alone in its efforts to regulate AI. US President Joe Biden recently issued an executive order on AI safety standards. China also has regulations related to AI, effective from last August.

Source: baochinhphu.vn