In recent months, a series of high‐profile controversies has highlighted growing tension between technology platforms, content creators, and regulators over how artificial intelligence is trained and used.
Previously, LinkedIn admitted sharing private messages to train its models, Meta was sued in France for using copyright‐protected content without authorisation, and OpenAI is paradoxically urging US lawmakers to expand "fair use" for AI. Meanwhile, several leading publishers, among them The New York Times and The Guardian, launched a "Support Responsible AI" campaign, accusing Big Tech of profiting from unlicensed journalism.
Against this backdrop, the UK’s Data (Use and Access) Bill, which would allow AI firms to use all published content unless rights-holders opt out, has stalled in Parliament, with over 300 peers demanding AI companies disclose and license any copyrighted material they employ. Baroness Beeban Kidron has proposed an amendment requiring the Secretary of State for Science, Innovation, and Technology to report on the bill’s impact on the creative industries three months after it becomes law, unless the bill is amended first.
Supporters of the government’s position, including Sir Nick Clegg, former president of global affairs at Meta, argue that requiring individual licences from all rights-holders would impose an unmanageable administrative burden and could deter AI investment in the UK. They say a broad opt-out regime is necessary for innovation.
Opponents counter that unlicensed use of creative content risks harming writers, musicians, filmmakers, and other creators by allowing AI tools to produce similar work without compensation. Prominent artists such as Sir Elton John, Sir Paul McCartney, and Dua Lipa have warned that unrestricted access to their work could undermine their income and legacy.
The debate traces back to early AI development, when large technology firms scraped substantial amounts of text, images, and other media from the internet without paying rights-holders. In 2022, the Ninth Circuit Court of Appeals held that "data scraping" does not constitute hacking, meaning it was not illegal to harvest publicly available website content so long as no defensive measures were bypassed. Those scraped datasets were then used to train generative models capable of producing text, images, and music in the styles of established creators.
The Department for Science, Innovation, and Technology is conducting a broader consultation and will not amend the bill unless changes are deemed satisfactory for creators. If the bill cannot be agreed upon, it risks being withdrawn.
The current impasse reflects a wider tension between supporting AI-driven innovation and protecting the economic interests of the creative industries. Until both Houses reach a compromise or one side concedes, the bill’s future remains uncertain.
3 Comments - Add comment