AI Copyright Changes

Kindus discussed AI texts and copyright in 2023. This is a rapidly evolving theme both in the nature of the ability of machines to create text and of other machines to identify content as created by humans or machines.

The UK government published a response to its AI opportunities action plan in January 2025.  Details are vague but they are clearly want to increase the UK involvement in AI developments. Initial steps being to set out a long term plan by Spring 2015 and to expand the government’s own AI capacity.

The copyright issues stem from the source material being used to create AI content. Much of this will be copyright but the AI engines may not correctly attribute it. There is also the debate as to who owns the newly created material. A report issued by the US Copyright Office in January 2025 concludes that:

‘… in many circumstances these outputs will be copyrightable in whole or in part—where AI is used as a tool, and where a human has been able to determine the expressive elements they contain. Prompts alone, however, at this stage are unlikely to satisfy those requirements’.

This provides an avenue for the ‘author’ of an AI generated work to claim copyright if they have had some expressive input into the creation of that work. It is not clear from what point this input moves from simple prompts to a more refined manipulation of the AI output.

In the UK the government published a consultation paper on ‘Copyright and AI’ in December 2024. This notes that AI created work is already protected by copyright law in the UK:

‘… section 9(3) of the Copyright, Designs and Patents Act 1988 provides specific protection for “computer-generated works” (CGWs). This protection, which lasts for 50 years, applies to a literary, dramatic, musical, or artistic work which is “generated by computer in circumstances such that there is no human author” of the work. The author of such a work is the person “by whom the arrangements necessary for the creation of the work are undertaken”. In the case of a general-purpose AI which generates output in response to a user prompt, the “author” will usually be person who inputted the prompt.’

This ruling could change depending on the outcome of the consultation paper but it appears that even entering simple prompts will confer copyright to that ‘author’.

These rulings impact on the difficult ground between copyright and intellectual property. If inputting prompts into an AI engine allows the ‘author’ to own a piece of work it may not mean that they have created something new that they can benefit from. A simple example would be a student essay relying heavily on AI. If the result is largely a copy of existing data and conclusions then it would be poorly graded or even rejected. If particularly clever or insightful prompts had drawn some innovative conclusions then the role of an AI engine would be more of a layout and grammar tool. A 2023 study published in the International Journal for Educational Integrity shows a variable performance within detection tools for identifying AI created content. ChatGPT4 output was harder to identify than ChatGPT 3.5 with some human created content being falsely identified as AI created. The study authors admit that using more and better detailed prompts from the ‘author’ should result in better AI output. As AI continues to evolve it could become very hard to identify and intellectual property rights would hinge on the output itself not the means used to create it.