Timing is everything in politics: when to call that election, when to U-turn on an unpopular policy, and when the moment’s right to spend more time with the family. It also matters to those trying to influence politics. You can have the best ideas in the world – or the worst – but you need to wait for a window of opportunity. When it comes to AI and copyright, that moment is now.
We’ve been talking about this for a fair while, and back in January last year we set out the challenges in our ‘Ronseal’ paper: Can the UK become competitive on text-and-data mining for AI? While we didn’t take an exact line on the solution, we wanted to draw attention to the fact that other countries were moving faster than the United Kingdom in securing the legal foundations for training AI models, with Japan, Singapore and the European Union each providing competitive models for the UK to consider.
Over recent months, this issue has become front-page news thanks to artists like Elton John, Paul McCartney and Dua Lipa adding their celebrated voices to the debate. Sadly, I don’t think this thorny policy is going to be solved by even the genius behind Yesterday – anyone who has a simple solution to what’s going on is either deluding themselves or lying to you.
As a new survey of 500 UK-based AI developers and investors revealed this week, there is a near-unanimous reliance on text and data mining (TDM) – the process of using automated techniques to train AI models. Of those surveyed, 99% said TDM is essential, while 76% said the absence of a copyright exemption for TDM – like that in the US, EU, or Japan – would make the UK less attractive for AI investment
For our part, we have a modest recommendation: listen to experts, including entrepreneurs. This is actually the Government line, as Ministers and officials pore over the 11,000 responses to a consultation they launched on copyright and AI in December 2024.
However, this process could all be derailed. On the back of the Data (Use and Access) Bill – a separate piece of legislation which addresses a broad range of priorities related to data – Peers have proposed adding changes to that Bill regarding how UK copyright law is applied to AI services.
Among other things, it would require developers to disclose comprehensive information regarding all text and data used in the pre-training, training and fine-tuning of general purpose AI on a monthly basis.
These proposals have been introduced outside the process of consultation that was launched precisely to inform the Government’s approach on copyright and AI. It’s a divisive issue, but surely we can all agree that any proposals with significant implications for businesses and innovation should be considered through proper channels. What’s the point of asking industry for their views if they are just going to be ignored?
Now is the time to act. It may seem perverse to dynamic founders who can pivot on a sixpence, but Westminster is a place where inertia too often rules the roost. Once the legislative train leaves the station, it can become impossible to switch tracks – even when many on board know their going the wrong way.
If you’re a founder, investor or just an interested party, drop me an email to find out how you could get involved.
Take a Punt
On Wednesday, we are hosting our second Young Entrepreneurs Forum Meetup in Cambridge. If you are a young founder – or wannabe founder – in or around the area, it would be great to see you there. (If you know of any young founders who you think should know, either pass this on or tag them in this LinkedIn post, which also helps get the word out either wider.)
The Young Entrepreneurs Forum is an international project from The Entrepreneurs Network. It aims to connect and empower young entrepreneurs who are building incredible things across the world, and to influence governments to create better enabling environments for young entrepreneurs.