EU moves closer to enact world's first law regulating AI systems

European Union lawmakers have pushed the bloc closer to passing one of the world's first laws regulating artificial intelligence systems like ChatGPT by backing a key text that forms the basis of a future law.

After months of heated discussions between lawmakers to support a compromise text, MEPs voted overwhelmingly to kickstart negotiations with the EU's 27 member countries for talks on Wednesday.

The EU insists that the law will foster AI innovation while also protecting against dangers the technology poses to people, and that once passed, it will set the "global standard". But the bloc could face pushback from big tech companies.

US-based OpenAI, which created ChatGPT, has already warned that depending on the content of the law it could be forced to withdraw from the EU.

While the bloc first proposed such a law in 2021, the draft rules took on greater urgency when ChatGPT exploded onto the scene last year, showing off AI's dizzying advances and possible risks.

Brussels is particularly concerned by deepfakes, AI-generated images and audio that can be stunningly lifelike, and how the technology can be used for disinformation, less than a year before European-wide elections.

Levels of risk

The law would regulate AI according to the level of risk: the higher the risk to individuals' rights, health, safety, or environment, the greater the systems' obligations.

The EU's proposed high-risk list includes AI in critical infrastructure, education, human resources, public order, and migration management.

The parliament's text differs from that proposed by the European Commission, the EU's executive arm, in 2021.

Lawmakers proposed bans on AI systems that use biometric surveillance including live use of facial recognition and so-called predictive policing. But the commission wants to allow such technology to be used in exceptional circumstances by security forces against crime and terrorism.

Commission Vice President Margrethe Vestager noted the parliament's approach and how it is pitted against "a slightly more pragmatic or security-oriented approach".

The text also calls for special requirements on generative AI systems, those such as ChatGPT and DALL-E capable of producing text, images, code, audio and other media, that include informing users that a machine, not a human, produced the content.

Brussels wants the final law to be approved by the end of the year. Even if that ambitious target is achieved, the law would not come into force until 2026 at the earliest, forcing the EU to push for a voluntary interim pact with tech companies.

The legislation builds on the EU's already formidable legal arsenal against technology companies, which includes two major laws to ensure social media platforms follow the bloc's rules.

However, it has not been welcomed by everyone.

The CCIA, a European industry lobby group representing major tech companies, warned some of the MEPs' changes "are likely to overburden European AI developers with excessively prescriptive rules, ultimately slowing down innovation".