Government plans to allow big tech firms to ignore traditional copyright rules when training their AI systems
Share:
The Government is planning to allow big tech firms to ignore traditional copyright rules when training their AI systems. Ministers are instead proposing that creative organisations, including music businesses and publishers, will have to opt-out if they do not want to have their work exploited by technology companies.
![[Lisa Nandy (pictured), Secretary of State for Culture, Media and Sport, said musicians, writers and artists should have the ability to know and control how their content is used by AI firms]](https://i.dailymail.co.uk/1s/2024/12/17/11/93229345-14201291-image-m-5_1734436318083.jpg)
There is huge controversy over the way tech companies use content from the UK to feed their AI systems, without paying or crediting these firms. The new consultation is proposing new requirements for AI model developers to be 'more transparent about their model training datasets'. This could see these companies made to provide more information about 'what content they have used to train their models'.
![[Secretary of State for Science, Innovation and Technology, Peter Kyle (pictured), said: 'The UK has an incredibly rich and diverse cultural sector and a groundbreaking tech sector which is pushing the boundaries of AI']](https://i.dailymail.co.uk/1s/2024/12/17/11/93229457-14201291-image-m-8_1734436501378.jpg)
But the proposals are likely to be controversial as they place the responsibility on the creators of the content to take measures to protect their copyright. There are also likely to be serious concerns about the government's ability to create a robust enforcement regime that would take tough action against AI companies that break the rules.
The proposals also risk putting smaller companies at a disadvantage because they would have to spend the time and potentially money to 'reserve their rights' rather than automatically having their copyright protected, as at present. Added to this are concerns the new rules will not be applied retrospectively to the massive amounts of content that has already been ingested into training AI systems without permission.
