UK government restarts review of AI training and copyright rules

In short: The UK government has started a new review of how copyright law should apply when AI systems are trained on books, music, and other creative work.

What happened

On January 14, 2026, the UK government announced a major review of AI training and copyright laws. Ministers described it as a “reset moment” after strong criticism from publishers, musicians, artists, and other creative groups.

The reset follows an earlier proposal that would have created a new legal exception for “text and data mining.” This is the automated copying and scanning of large amounts of text or images (like a robot reading millions of pages) to help train AI. Under that earlier plan, AI developers could have used copyrighted work unless the creator actively opted out.

That opt-out approach was unpopular in the government’s own consultation. Only about 3% of respondents supported it, according to the reporting. Creative industry groups said it would force creators to police the internet for misuse, while AI companies would get broad access without paying for licenses.

The government says the new review, due to report in March 2026, will look for a balance. It will consider how creators can get recognition, control, and fair payment when their work is used for AI training, while still allowing AI development.

Why it matters

This affects everyday things like how writers, photographers, and musicians get paid, and how quickly new AI products show up in the UK. It also matters for ongoing court cases, including a dispute between Getty Images and Stability AI, which is testing what happens when an AI model trained abroad is sold or used in the UK.

Source: Financial Times