Companies launching generative AI tools, such as ChatGPT, will have to disclose any copyrighted material used for training their AI systems, proposed an agreement in the European Union, according to a Reuters report on April 27. “Companies deploying generative AI tools, such as ChatGPT or image generator Midjourney, will also have to disclose any copyrighted material used to develop their systems,” the report added. It also states that some committee members initially recommended banning the use of copyrighted material for training generative AI tools altogether, but the proposal was “abandoned in favour of a transparency requirement”. According to an anonymous source speaking to Reuters, members of the European Parliament are all set to work out the final details of the bill for AI regulations, which has been in the works for quite a while now. Why it matters: To what extent can existing works be used for machine learning purposes without copyright infringement, is the fundamental question here. Regulators in the US, China, and European Union are at the forefront of devising AI regulation rules and mechanisms, but the EU has taken a step ahead in addressing the copyright infringement issue specifically. The upcoming EU rules may pave the way for other countries to modify their copyright laws in order to address AI-specific issues. Will disclosure mandate solve the issue? The non-licensed use of data available on the internet for training generative AI systems by companies is a critical point of contention between AI developers and artists, publishers, among others.…
