Mark Twain famously quipped that he didn’t have time to write a short letter, so wrote a long one instead. The modern twist is the observation that most books should be articles, most articles blogs, and most blog posts tweets. (I’m not sure what most newsletters should be.)
Taking this advice to heart, this year we’re launching a new way to publish research. ‘Explainers’ allow us to really hone in on policy issues impacting entrepreneurs without the rigmarole of writing a full report.
Our first explainer came out today. Dr Anton Howes, our Head of Innovation Research, has written an easily digestible 1,300 words on how UK copyright is impacting the way founders train AI. Anton is reporting from the coalface, having been involved in countless meetings and negotiations on the topic with the government, regulators, AI companies and rightsholders.
As his thread on X (Twitter) explains (likes and shares are always appreciated): “It’s become a fraught issue lately. In 2022 the UK government announced it would implement an exception from copyright for all copying of material made for the purposes of text-and-data mining. But this provoked a backlash from the creative industries, and in 2023 the government dropped its proposals. Since then, it has been trying to host negotiations for a voluntary code of conduct for AI companies and copyright owners.”
The ongoing lack of clarity has been disastrous for AI startups working in this space, and confusing for creatives who don’t know where they stand.
It’s a prickly policy area, but Anton turns to other countries for inspiration. As far back as 2018, Japan created a broad exception for text-and-data mining of lawfully accessed material, on the condition that copies made are read only by machines, and not humans.
In Singapore, reforms enacted in 2021 created a blanket exception for text-and-data mining of lawfully accessed material, with a similar proviso that the copies must be made only to be read by computers, not humans. While in Israel, the Ministry of Justice has attempted to clarify things by issuing an opinion that copyright law does not prevent the mining of lawfully accessed copyrighted material.
The EU’s 2019 Copyright Directive extended the exception to mining for both commercial and non-commercial uses, but it also allows copyright owners to explicitly opt- or contract-out of having their material being used for training AI. As Anton concludes in his thread: “Funnily enough, this is very similar to an option that was floated in the UK before the government made its 2022 announcement that it had to u-turn on!”
This sort of uncertainty impacts investment and jobs. It’s time for clarity.
Got Form
While important for UK economic growth, the odds are high that text-and-data mining isn’t a key issue for you. So what’s your small or big ask?
If we think what you’re suggesting is a step towards making the UK the best place in the world to start and grow a business, we will amplify your voice. It’s why we exist.
Whether it’s for our explainers or for our upcoming manifesto report now is the time to let us know what you – or the businesses you work with – need to succeed. I put out requests for your ideas quite regularly, but that’s because we always get interesting responses. So here is a Google form to make the whole process easier.
As our Head of Research Eamonn Ives wrote in his latest APPG newsletter: “In an ideal world, the trials and tribulations of government affairs wouldn’t be a concern for business owners – who we would rather were able to focus their energies on improving their products, upskilling their staff and so on. But the truth is, politics does matter. And it matters that entrepreneurs’ voices are heard at the volume they deserve.”
Sign up to my Friday Newsletter here.