The Government has scrapped plans to make publishers ‘opt out’ if they want to stop their works being scraped by LLMs.
Ministers had intended to change copyright laws to allow companies to train their AI models on others’ intellectual property, unless permission was actively revoked.
But after publishing a report and impact assessment on copyright and AI on 18 March, the Secretary of State for Science, Innovation and Technology, Liz Kendall, said that the government has abandoned this intention, and that it ‘no longer has a preferred option’ for how to regulate AI.
The announcement came two weeks after the Professional Publishers Association (PPA) warned that the government’s proposed opt-out mechanism for text and data mining by LLMs would result in publishers being unwilling to use the protections offered to them for fear of ‘punishment’ by Google.
Sajeeda Merali, CEO of the PPA, responded positively to the statement, saying that the Government showed that ‘it is listening to the consistent feedback from across the sector’.
Merali praised the acknowledgement that the opt-out model is ‘flawed’ and said that dropping that option ‘would represent important progress for publishers and the trusted editorial brands our members represent’.