Met Office data and legal documents from the National Archives could be used by artificial intelligence systems as the UK government pushes ahead with plans to employ nationally owned content in AI tools.
The government is providing funding to researchers to test how Met Office materials could be used by technology, such as to help agencies and councils know when to buy more road grit. Another project will explore whether legal data from the National Archives – the UK’s repository for official documents – can provide legal support to medium and small-sized businesses.
The government has also announced plans to license material from national institutions such as the National History Museum and the National Library of Scotland for AI development.
Digital Government and Data Minister Iain Murray said the National Archives plan looked like “smart use of the public sector”.
“Whether you’re a café owner trying to understand employment law or a shopkeeper trying to understand health and safety requirements, time spent battling legal jargon is time spent away from customers,” he said.
“By exploring how official legal data from the National Archives can be made AI-ready, we can help small businesses get quick, reliable answers to common questions – freeing up evenings and weekends and giving them more confidence to invest and grow.”
Data is a key component of AI models, the underlying technology for tools like ChatGPT, and has become a battleground between the UK’s creative industries and the government over proposals to change copyright law.
Ministers have been criticized for proposing that AI companies should be allowed to use copyright-protected works without the artists’ permission first, unless the artists indicate they want to “opt out”. Liz Kendall, the technology secretary, said this month that the government was seeking a “reset” on these plans through an official review to be published in March.
However, the government has clarified that nationally owned data can also be used as part of its policy to drive AI deployment and development.
The government’s AI action plan, launched last year, proposed creating a national data library that would provide state-controlled data to researchers and companies. Keir Starmer indicated that the vast stores of health data controlled by the NHS would be candidates for this library, although the government has also stressed that privacy, ethics and data security issues will be taken into account when collecting these datasets.
The action plan also proposes creating a cultural data library that would be a “commercial proposition”, meaning the government could expect to make money from “copyright-granted” data. It can hold data from institutions like BBC, British Library, Natural History Museum and National Archives.
The government will on Monday put forward plans detailing the data to be introduced under its “creative content exchange”. The exchange, which launched in June, will be a marketplace to “enable the sale, purchase, license and access of digitized cultural and creative assets” so that they can be “licensed at scale”.
Under a pilot scheme, researchers will explore how to introduce data from institutions including the Imperial War Museum; National Library of Scotland; the natural History Museum; National Archives; University of Oxford; Royal Botanic Gardens, Kew; Science Museum Group and the Victoria and Albert Museum.
The government said the scheme will explore how their digital content can be used by consumers, technology companies and AI developers, “while respecting the rights of creators and copyright owners”, with a pilot platform to be launched in the summer.
