“Both sides of this debate benefit from copyright law,” said a joint response to the UK government’s AI consultation from the UK parliament’s cross-party culture, media and sport committee and the science, innovation and technology committee.
The government is proposing a new exemption in copyright law that would allow tech companies to train their AI models on creative works including films, TV shows and audio recordings without permission, unless creators actively opt out – akin to the European Union’s approach.
The letter was addressed to Peter Kyle, secretary of state for science, innovation and technology and Lisa Nandy, secretary of state for culture, media and and sport. It was signed by Chi Onwurah, chair of the science, innovation and technology committee and Caroline Dinenage, chair of the culture, media and sport committee.
Copyright law and fair renumeration for creative work, the letter stated, is imperative for “incentivising further creativity and innovation”.
The committees issued a warning to tech companies themselves over meddling with the UK’s copyright framework, pointing out that original non-literary written works, such as software, web content and databases, are not protected by patents under UK law, but by copyright.
“Any industry, particularly a new and emerging one, should take a considered approach when advocating for the reform of frameworks that may in fact be protecting their economic interests,” said the letter.
“The case of DeepSeek, which OpenAI alleges breached ChatGPT’s terms of service by copying outputs to train its model, illustrates these potential risks. Additionally, we note that other key jurisdictions, such as the United States and European Union, have not settled this issue, despite the government’s statement that AI developers seek countries with ‘clearer or more permissive rules’.
“In the US, where copyright law contains ‘fair use’ provisions, almost all of the major generative AI developers are collectively facing dozens of lawsuits, while in the EU, whose approach the government’s preferred option broadly emulates, the German collecting society GEMA has filed lawsuits against OpenAI and music generator tool Suno in recent months.”
The committees have made four key recommendations to government: first, practical measures should be introduced to make sure companies are transparent on how AI data is being trained, regardless of the government’s approach to copyright law; second, the government should not move forward without a technical solution that is implementable and accessible to all, and that that may not happen as quickly as government ministers are hoping; third, the government should publish a full impact assessment for each option set out in the consultation; and fourth, robust mechanisms must be put in place to enforce compliance and give the opportunity for creatives to take issue with their copyrighted work being reproduced, and not only if they have the funds to take it to court.
The letter also expressed disappointment that, during a joint parliamentary hearing on the subject of AI and copyright earlier this year, both of the two leading AI developers, Google and OpenAI, declined to take part. “We are of the view that, given the extent to which AI has been embraced by successive Conservative and Labour governments, the developers of these models should take a more constructive, transparent approach to engaging with parliamentary scrutiny.”
The government’s consultation on AI and copyright closed on February 25.
No comments yet