AFP – European Union lawmakers on Wednesday pushed closer to passing one of the world’s first laws governing artificial intelligence systems like ChatGPT, backing a key text that forms the basis of future legislation.
While the EU first proposed such a law in 2021, last year the ChatGPT overshadowed the draft rules by showing AI’s dizzying advances and potential risks.
The EU says that once passed, its legislation will set a “global standard”, but the bloc could face pushback from big tech companies.
US-based OpenAI, which created ChatGPT, has already warned that it could be forced out of the European Union depending on the content of the law.
While AI proponents praise the technology for how it will transform society, including work, health care and creative pursuits, others worry about its potential to undermine democracy.
There are awe-inspiring examples of AI uses including the “last Beatles record” created using the technology, while ChatGPT has passed business, law and medical exams.
EU lawmakers voted overwhelmingly to kickstart talks with the bloc’s 27 member states after months of heated discussions among lawmakers to endorse the agreement text and talks were to begin later Wednesday.
The EU insists that the law will spur AI innovation while also protecting people from the dangers posed by the technology.
Brussels is particularly concerned with deepfakes – AI-generated images and audio that can sound surprisingly lifelike – and how the technology can be used for disinformation, with European-wide elections less than a year away Earlier
The law will regulate AI according to the level of risk: the greater the risk to the rights or health of individuals, for example, the greater the system’s obligations.
Brussels wants the final law to be approved by the end of the year.
EU lawmakers vote on the Artificial Intelligence Act in the European Parliament on June 14, 2023 in Strasbourg, eastern France. (AP Photo/Jean-Francois Badias)
Even if that ambitious target is achieved, the law won’t come into force until 2026 at the earliest, forcing the EU to push for voluntary interim agreements with tech companies.
Parliament’s text differs from the text proposed in 2021 by the European Commission, the EU’s executive arm.
Lawmakers proposed a ban on AI systems that use biometric surveillance, including facial recognition and live use of 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 Parliament’s approach and how it was pitted against a “slightly more pragmatic or security-oriented approach”.
Text also calls for special requirements on generative AI systems – such as ChatGPT and DAL-E capable of creating text, images, code, audio and other media – including informing users that a machine, not a human produces material.
As member states and parliament hammer out a deal in the coming months, officials said sanctions and generative AI will continue to dominate the talks.
The law builds on the EU’s already formidable legal arsenal against technology companies, including two key laws to compel social media platforms to comply with the block’s rules.
ChatGPT app on iPhone, May 18, 2023. (AP Photo/Richard Drew)
The EU’s proposed high risk list includes AI in critical infrastructure, education, human resources, public order and migration management.
Parliament added additional conditions before a high-risk classification could be met, including the potential to harm people’s health, safety, rights or the environment.
This has not been welcomed by all.
The CCIA, a European industry lobby group representing major tech companies, warned some lawmakers the changes “are likely to overburden European AI developers with overly prescriptive regulations, ultimately slowing innovation.”
Brando Benifie, one of the EU lawmakers, said on Wednesday, “We have made history today.”
They had defined the EU’s dialogue with the rest of the world on how to regulate AI, the lawmakers said.
There is also growing noise across the Atlantic to regulate AI, although Washington lags far behind the European Union in developing such rules.
