Key takeaways from Keir Starmer’s action plan for AI

1. A sovereign AI drive

Ministers want to ensure that the UK is able to produce its own AI models – the technology that underpins tools such as the ChatGPT chatbot. To that end, the plan recommends a 20-fold increase in the amount of AI computing power under public control by 2030. This will include building a brand new “supercomputer” at a location to be decided.

Then there is data, the core fuel of AI. If AI can be broadly described as computer systems performing tasks that typically require human intelligence, then they need data – from academic papers to classic novels and transcripts – to “learn” how to carry out those tasks.

The plan says the government should create a national data library, comprising state-controlled data, that will be made available to researchers and companies. At least five “high-impact” data sets will be made available, although the plan notes that privacy, ethics and data protection considerations will have to be taken into account. The prime minister signalled on Monday that the vast trove of health data controlled by the NHS will be a candidate for this library.

The plan also proposes a cultural data library, snappily called a “British media asset training data set”. This could hold data from institutions such as the BBC, the British Library, the Natural History Museum and the National Archives. The plan says this set will be “copyright-cleared” and makes clear the government should make money from it, calling it a “commercial proposition”.

2. Encouraging the private sector

The government will create “AI growth zones”. A streamlined planning process will help them host data centres – the central nervous system of AI technology – while they will also accelerate the provision of “clean” energy to power that infrastructure. The Oxfordshire-based Culham Science Centre, which is the UK Atomic Energy Authority’s headquarters, has been singled out for a potential dry run for a growth zone.

Recruitment of elite AI talent is also mentioned. The plan recommends establishing a headhunting unit that will bring in top talent from abroad for the UK’s AI safety institute and publicly owned labs – but also UK-based companies. It also suggests shaking up work visas to help tech startups attract non-UK talent.

The plan also calls for “sector champions” to drive adoption of AI across the UK economy in key sectors like banking and the creative industries. The government should also help workers adapt to AI in terms of new skills and new jobs.

There is also the ambition to create “national champion” AI companies, with the plan proposing the creation of a new body called “UK Sovereign AI” to support entrepreneurs and help create the next Google DeepMind – a UK-founded world leader in AI research.

3. Embedding AI in the public sector

The plans calls on the government to “rapidly pilot” AI-powered services, saying this will drive better experiences and outcomes for citizens, as well as boosting productivity – a measure of economic performance. The best pilot schemes will then be deployed UK-wide.

Examples of public sector use of AI flagged by the government include helping teachers prepare for lessons and using AI-enhanced road cameras to spot potholes. The logic behind this drive is to improve public services but also to save money by automating tasks, which carries the implicit threat of job reductions across the public sector.

4. AI and energy

AI is an energy-intensive technology due to its use of electricity-guzzling data centres, which means there are clear environmental costs to increasing the UK’s role in driving the technology. According to the International Energy Agency, total global electricity consumption from data centres could reach the level of Japan’s energy intake by 2026.

The action plan refers to the AI growth zones using “clean power” and the government announced on Monday that it will set up a new AI energy council to make that happen. The body will be co-chaired by the technology secretary, Peter Kyle, and the energy secretary, Ed Miliband. The council will seek to accelerate investment in renewable energy for data centres as well as small nuclear reactors. Last year, Microsoft, Amazon and Google. who are big players in AI, all announced agreements to use nuclear energy.

5. AI and copyright

The plan makes a recommendation to reform the UK’s regime for copyright, which is a legal protection against someone’s work being used without permission. It says the UK’s framework for text and data-mining – ie, trawling the internet for copyrighted material and feeding it into AI models – should be “at least as competitive” as the EU’s. This would mean allowing the use of copyrighted material by AI firms provided they give copyright holders the ability to opt out of the trawling process. To campaigners for strong copyright protection, this is tantamount to trampling on the rights of creative professionals and publishers.

Ed Newton-Rex, the British composer and a prominent voice in the copyright debate, says the plan’s support for copyright reform is a “huge blow to the UK’s creative industries”.