The opt-out model would have flipped copyright’s default position. Under it, AI developers could have used any UK-created work for training unless the rights holder specifically said no. Rights holders, individual authors, journalists, musicians, and the organizations that represent them, would have carried the burden of protecting their own work from commercial AI use.
The UK government has decided not to proceed with that approach. The announcement came on March 18, 2026.
Bectu, the union representing workers in film, television, and media, confirmed the decision and welcomed it directly, citing the disproportionate burden the opt-out mechanism would have placed on freelancers and early-career workers. The National Union of Journalists also published a response to the government’s report on the same day.
The government’s decision followed a public consultation that produced overwhelming opposition to the opt-out model from the creative industries, according to coverage of the government’s report. Specific figures from the consultation haven’t been confirmed from verified sources, but the scale of opposition is described as decisive.
What the government has indicated is that it will explore alternative approaches, frameworks that address both the needs of AI developers and the rights of creators. That likely means some combination of licensing frameworks, transparency requirements for training data, and fair remuneration mechanisms. None of those alternatives have been confirmed in detail.
For AI companies relying on UK-sourced training data, the practical position has shifted. The permissive default they were counting on isn’t coming. For rights holders, the immediate win is real, though the replacement framework isn’t settled yet.