The UK government has reversed its plan to let AI models train on copyrighted material without paying creators.
This decision follows strong backlash from across the creative industries.
It now stands as a major win for artists, labels, publishers, and rights holders.
On 18 March, Culture Secretary Liz Kendall confirmed the shift in direction.
The government no longer supports the earlier proposal, which would have allowed AI companies to use licensed creative work by default.
Although rights holders could opt out, critics argued that the system would still enable large-scale use of protected content without proper consent.
The proposal quickly sparked a wide campaign.
Major figures in music, including Elton John, Dua Lipa, Björn Ulvaeus of ABBA, Max Richter, and Thom Yorke of Radiohead, pushed for change.
They warned that the plan could weaken copyright protections and harm creative industries.
Kendall explained how the government arrived at its decision.
She pointed to a long process of discussions and consultations.
Creatives, tech companies, unions, academics, and industry groups all took part.
Through these exchanges, the government gradually reshaped its stance.
She also noted that many countries, including the United States, the European Union, Australia, and India, face similar challenges.
For that reason, she stressed that finding the right balance will take time.
Read more: Bandcamp Bans AI-Generated Music to Protect Artists
The Reaction from Industry Leaders
Industry leaders have welcomed the move.
Mandy Hill of Cambridge University Press described the decision as a win against the interests of a small number of large corporations.
She also reinforced that current law already requires permission before using copyrighted material for AI training.
However, some concerns remain.
UK Music called the decision a major victory but urged the government to confirm that the proposal will not return in another form.
Similarly, the British Phonographic Industry expressed relief while warning against future exceptions, especially around text and data mining.
The group highlighted the potential for growth if licensing systems support both creativity and AI innovation.
Meanwhile, Tech UK pointed out the need for balance.
The group warned that overly strict rules could slow down the country’s fast-growing AI sector, which is expanding far quicker than the broader economy.
A recent report also noted that there is still no clear agreement on how to align the interests of AI developers and creative industries.
The urgency of the issue continues to grow. In 2024 alone, around 60 million people created music using AI tools.
Output has surged rapidly, with platforms like Suno generating volumes comparable to Spotify’s entire catalog on a monthly basis.
In a related move, OpenAI has explored new licensing models to generate revenue.
One option includes outcome-based royalties, where users pay based on the value of what AI produces.
Critics have challenged this idea, arguing that it may undervalue the original copyrighted works used to train these systems.
