Regulations Proposed By The EU Would Regulate Generative AI

After almost two years of work, the EU’s AI Act has reached the trilogue stage. The draft says that creative AI tools like ChatGPT must reveal the copy-protected materials they used to make their programs.

The goal of the act is to control new AI technology, which has been growing rapidly in recent months, with Alphabet creating a new division focused on AI and the UK government wanting to be at the front of the AI race.

Under the plans, AI tools will be put into four groups based on how dangerous they are thought to be: minimal, limited, high, and unacceptable. Concerns could include biometric monitoring, spreading false information, or using language that makes people feel bad about themselves.

In this draft, high-risk tools are not banned, but those who use them will have to be very open about it. By revealing their development materials, generative AI tool developers like Midjourney will be held to this transparency criterion.

Reuters says that this clause was added at the last minute, in the last two weeks. According to the source, some committee members first recommended preventing copyrighted material from training generative AI models, but this was abandoned in favor of openness.

“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.