Copyright and AI: What You Need to Know

By Peter Stead

You may have heard about a Government policy move to let AI Companies train on copyrighted content without permission from the copyright holder. If you are mostly confused about this issue, you are not alone. A few years ago, I myself knew very little about AI – or copyright, for that matter. I still have much to learn, but I made it my business to gather the relevant basics, because of how deeply concerning this move is.

Let’s start at the beginning.

At present, there is already a legal AI Exception to copyright law that allows for training on copyrighted material without permission, if the purpose is non-commercial.

Back in 2022, the then Conservative Government announced plans to expand this Exception in law to encompass any purpose, including, of course, commercial. This is despite the consultation that they conducted at the time showing strong opposition from the Creative Industries, because, of course, it would have greatly prejudiced the earning potential of their Intellectual Property (IP) if it could have been copied at will. Following a backlash, the plans were dropped.

Or were they?

In 2024, the new Labour Government announced they wished to pursue the Exception again, but modifying to add the right for you to opt out of having your works trained on for AI. The Creative Industries were, again, quick to point out the obvious impracticality of this approach, considering that it would mean having to track down every piece of your work on the internet in order to exercise this opt-out. What we already have is much better, because the onus is on AI Companies to get your permission.

Some key points to understand here:

Even with an opt-out, the move would have been illegal under international law and you can read Nicholas Caddick KC’s very clear legal opinion here. Big questions also arise around whether the move would be compliant the Human Rights Act 1998, with which all UK law is meant to follow. Article 4 of the HMA deals with freedom from slavery and forced labour, and Protocol 1, Article 1 deals with the right to peaceful enjoyment of your property, and yes, this includes intellectual property.

At the end of this new consultation, which received over 13,000 responses, the Government stated that this is no longer their preferred approach. At the time of writing, we do not know what the Government’s next move will be, but we do know they are blocking Baroness Kidron’s attempts to force AI Companies to be transparent about their training data.

Wider Context

I have written above about copyright, especially through the lens of the 2.4M-strong Creative Industries that contribute £124.6Bn, or 5.7% to the UK Economy. However, the real issue is broader than that. It is Human Rights. As I wrote above, it is difficult to see how the Government’s proposed move would not be in breach of two articles of the Human Rights Act 1998. Allow me to mention a third, Article 8, that confers a right to privacy, a basic concept which is also enshrined in law as the Data Protection Act 2018, which incorporates UK GDPR.

To try to start shedding light on this whole debacle, I made Freedom of Information Requests to the Intellectual Property Office regarding the very first Exception the Government pursued, and if you want a deeper dive into that part of this whole issue, you can read the first ICO decision notice here. In brief, I won some materials, while others were withheld. However, I am hoping to have the withheld materials released to me after I appeal to a judge-led panel of the First Tier Tribunal.

Final Thought

Although the destruction of copyright is egregious enough in its own right, this proposed move is also a savage assault on fundamental human rights. And in the final analysis, it is human rights that I and others are fighting to defend.