The EU’s AI Act has been pushed through to the trilogue stage after nearly two years of drafting. The draft includes requirements for generative AI tools, like ChatGPT, to disclose the copyrighted material used in developing their programmes.
The aim of the act is to regulate emerging AI technology, which we have seen booming recently with Alphabet forming new AI-focused division and the UK government wanting to lead the AI race.
Under the proposals, AI tools will be classified according to their perceived risk level: from minimal through to limited, high, and unacceptable. Areas of concern could include biometric surveillance, spreading misinformation or discriminatory language.
High-risk tools are not banned in this draft, but those using them will be required to have a high level of transparency. Those creating generative AI tools including image generators like Midjourney, will also be held to this transparency threshold by disclosing material used to develop their programmes.
According to Reuters, this provision was a late addition drawn up within the past two weeks. Some committee members initially proposed banning copyrighted material being used to train generative AI models altogether, the source said, but this was abandoned in favour of a transparency requirement.
"Against conservative wishes for more surveillance and leftist fantasies of over-regulation, Parliament found a solid compromise that would regulate AI proportionately, protect citizens' rights, as well as foster innovation and boost the economy," said Svenja Hahn, a European Parliament deputy.