Sir Keir Starmer isn’t usually one for grand visions. Yet, on a Monday at University College London, he took a different path, addressing a topic he feels could define our generation: artificial intelligence. Starmer sees the UK, once the cradle of pioneers like Babbage, Lovelace, and Turing, and the birthplace of the modern computer and internet, rising as an AI superpower.
It was a compelling narrative. Just days into his leadership, the Prime Minister enlisted Matt Clifford, a tech-savvy personality, to explore the nation’s AI potential. Clifford crafted a comprehensive 50-point AI Opportunities Action Plan, which Starmer wholeheartedly embraced, vowing full governmental backing. Clifford was also appointed Starmer’s AI Opportunities Adviser, running the plan’s implementation and reporting directly back to him. It’s likely only a matter of time before he earns the moniker "the UK’s AI tsar."
Clifford’s involvement is both anticipated and intriguing. He’s no stranger to governmental circles; Rishi Sunak had previously tapped him to orchestrate the AI Safety Summit and establish the UK’s AI Safety Unit. The puzzling aspect lies in his significant financial success in the tech industry, with a lengthy list of external interests. Insiders at tech firms and media outlets expressed concerns to the Financial Times about Clifford’s growing sway in shaping AI policies, given his background in building a successful global investment firm.
"He’s undeniably capable," noted Damian Collins, a former Tory technology minister, adding uncertainty about how interests in the field are being represented. With AI viewed as transformative by Starmer, it seems odd that his key advisor is deeply involved in the game.
Collins pointed to the contentious issue of copyright violations by tech companies training AI models using others’ creative works without due credit or payment. Fresh details from a US court case revealed that Meta’s Llama AI training dataset included a massive trove of illegally sourced books.
The 24th recommendation of Clifford’s plan proposes reform of the UK’s text and data-mining policy, citing a lack of clarity around intellectual property (IP) which hinders innovation and the growth of creative industries. This claim has infuriated parts of those sectors. The Creative Rights in AI Coalition insists there’s no legal ambiguity: UK copyright law prohibits commercial text and data mining without proper licensing. The question remains who exploited the UK’s creative assets without consent.
Despite the controversy, Clifford’s plan seems well-rounded, albeit costly: building AI computing infrastructure; expanding research capacities in universities; training countless new AI specialists; nurturing public-private partnerships to enhance the UK’s AI "frontier" involvement; and ensuring rigorous ethical and technical standards in AI evolution.
This approach is a stark departure from the hollow rhetoric of "Global Britain" from previous leadership eras. The plan’s ambition, to position the UK as an AI creator, not just a user, acknowledges the nation’s untapped potential yet candidly recognizes the lack of resources to realize it. Achieving this ambition requires confronting a couple of harsh truths.
First, the AI arena is predominantly governed by a few colossal corporations, none headquartered in the UK. Their dominance stems from not only capital and talent but also their vast data centers. Any nation wishing to play in this field needs to navigate relationships with these giants carefully.
In this regard, the UK government should step up its game. The current deferential stance, evident when technology secretary Peter Kyle suggested a “sense of humility” in engaging with tech giants rather than resorting to legislative threats, needs rethinking. Essentially, they suggest treating these corporations as sovereign states. As the saying goes, appeasing them could be as futile as hoping a crocodile will spare you if you’re polite.
The second truth is that despite AI’s power, economists like Nobel laureate Daron Acemoglu predict its overall economic impact might be less than tech enthusiasts expect, at least initially. Additionally, as economist Robert Gordon once highlighted, transformative technologies take time to significantly impact economies. The message for the Prime Minister is clear: becoming an "AI superpower" won’t happen overnight – it’s a long-term goal, possibly spanning multiple electoral terms.
John Naughton serves as the chair of the advisory board of the Minderoo Centre for Technology and Democracy at Cambridge.
