AI and copyright

I am not at all against Artificial Intelligence. It is and will be an extraordinarily powerful tool in science and medicine and will certainly have limitless uses in all sorts of unimaginable ways. What I am talking about here is generative AI. Specifically generative AI platforms that are trained on unauthorised copyrighted images, text and video, which is all of them.

I asked the AI image generator MidJourney to create ‘a squirrel in the style of Chris Haughton’ At first it generated a photorealistic squirrel wearing my glasses. But after I clarified that I wanted images in my art style rather than my style of dress it came back with images that certainly had my colours, shapes and backgrounds. It looked like art I might draw on a bad day. But it was my work nonetheless. And that is the thing. It was my work.

I then asked ChatGPT to create a story in the style of Chris Haughton. It came back with language and content that was familiar about a lost animal. Again, it had been clearly referencing my stories. When I looked to see if my images were in the training set at ‘Have I been Trained?’ my images were everywhere, thousands of them. Many in duplicate. One of my images, a book cover, appears duplicated in the set 88 times.

These AI tools were trained on hundreds of millions of images. And on billions of words. The content they used were created by countless artists, authors, translators and journalists. The UK government is arguing that ‘there is uncertainty about copyright’. But there is absolutely no uncertainty about copyright: AI companies took copyrighted images and texts, which in many cases took years to make, without asking, and have used them for commercial purposes. Under copyright law this is illegal.

The UK government is currently planning to change this. We have a very short window of opportunity to object, their deadline is the 25th February. The government is proposing to add an ‘exception’ to copyright law allowing AI developers to scrape your work from the internet and use for training AI and creators will need to manually ‘OPT-OUT’ of this exception. The problem with ‘opting out’ is it in practical terms completely impossible. Ed Newton-Rex very clearly summarises the impossibility of opt-out here

A situation where everyone who does valuable work is rewarded for it is a good economy. Creators who make valuable work will be remunerated for it and so are encouraged to do more, and everyone benefits from this. Many creators would be very happy for AIs to use their work if they got paid for it. Just like I am happy for my publisher to publish my books because they pay me for it. I would not be so happy if I created a picture book and it was published worldwide and the publisher became a billionaire and I got no money from it. That was pretty much the situation in the early 1700’s when copyright as we know it came in to force. That is why copyright exists. We realised we need it if we want a healthy publishing industry and we want quality books.

The situation that the UK government is proposing is the opposite of that. They are proposing AI can use our images without paying us. Millions of creative people doing work which is then taken for free by a handful of companies. Over 2.4 million people work in the creative industries in the UK. A handful of companies, mostly from the US not the UK, have trained their AIs on billions of works made by these creators for free. The value of their AI is only as good as the works they are trained on. Their value is our value. Yet we get paid nothing.

This is yet another example of the larger picture of the world today. In 2000, when Google launched their advertising model, AdWords, they were transformed from a cash-strapped student startup into one of the richest companies in the world. All of the other platforms naturally followed this model. But this meant that these companies had morphed from open platforms for sharing, to extractive surveillance tools that work against us.

Everyone that uses services like Google or social media is giving away their personal information for free. This information is sold by Google, Meta (Facebook and Instagram) and the other Big Tech platforms for billions of dollars. We give the data for free to Big Tech and they sell it for billions. That is the system. This, no matter what way you look at it, is not a fair deal. If someone had proposed this system at the start nobody would have agreed to it.

Big Tech’s power is truly extraordinary. Some have put it as much as 99% of spending on digital ads goes to either Google or Meta/Facebook. At $1.2 trillion these two companies are worth more, than every single newspaper and publisher in the world put together.* This puts them in an very, very powerful position.

Naturally, they want this unequal system to continue. That is why they are today they are some of the largest donors to the US, UK and other governments. Governments are becoming corporatised by Big Tech. We saw in January how Mark Zuckerberg, Elon Musk, Jeff Bezos and other Big Tech figures attended Donald Trump’s inauguration together. They were there because they were all amongst Trump’s largest donors.

The UK has a very similar pattern of political corruption. Usually Labour’s largest donations are from the Unions but Keir Starmer’s government has seen an increase in corporate donations, mostly from Big Tech, which has weakened the Union’s strength. Labour has also appointed several big tech insiders to advisory roles. For instance, Microsoft UK’s CEO, Clare Barclay, now leads the government’s Industrial Strategy Advisory Council. Lord Vallance, a former scientific adviser and executive at Google-owned Isomorphic Labs, was appointed as Science Minister, raising concerns about AI regulation and copyright laws.

We need to urgently examine this. These extractive corporations are now working hand in hand with government. It is this that is causing the increasingly rightward shift around the globe. Unless this is not tackled it will only move further rightward. We have seen this happen before. In the nineteenth and early twentieth centuries extractive corporations and government also worked in unison.

Armies captured territories, legal instruments were set in place, and extractive and exploitative corporations followed. These Imperial projects grew and grew until they reached their pinnacle with fascism. As Mussolini himself said “Fascism should more appropriately be called Corporatism because it is a merger of state and corporate power.”

The collusion of Big Tech with government is strikingly similar to how these colonial projects operated. It is all about extraction and exploitation. This is not a matter of artists and people in the cultural sector worried about their livelihoods, it is a matter that should be of utmost concern to us all.

If this change in copyright law does go through it will be almost impossible to roll back. Please share.

I was drawn into this debate thanks to the brilliant Association of Illustrators. They are the closest thing illustrators have to a union. They are currently fighting alongside representatives from all the creative fields, authors, film-makers, journalist, musicians across Britain to stop this change. Thank you Derek Brazell for the the phone calls and messages and help with crafting this text. Thanks too to Momoko Abe, Ged Adamson and Simona Ciraolo for their activism around this.