If you can’t join them, don’t beat them.
A leading figure in the development of generative artificial intelligence has warned that the UK should stop trying to compete directly with the major players dominating the global AI landscape.
Llion Jones, a Welsh computer scientist who helped create the Transformer architecture that underpins modern AI systems, said the UK must be “brave” and pursue alternative approaches if it wants to remain relevant.
Jones, who co-authored the landmark 2017 paper, “Attention Is All You Need,” while at Google, argued that the scale and resources of the United States and China have already placed them far ahead in the competition to build the largest and most powerful AI models.
“I think a country like Wales, or even Britain in general, is not going to beat the hyper-scalers, right? They’re not going to beat America and China in scaling up these AIs,” he said in a BBC news report. “It doesn’t make any sense to be a part of that race, because you’re not going to win.”
His comments come as governments worldwide assess how to position themselves amid the rapid acceleration of AI capability, which is reshaping industries, public services, and geopolitical relationships.
A call for more distinctive AI research
Jones, who grew up in Bangor and Abergynolwyn in Wales, now serves as chief technology officer at Sakana AI, the Tokyo-based company he co-founded with another former Google researcher. Unlike most major tech firms, which are focused on commercialising large language models, his company explores speculative and unconventional forms of AI research.
“I do genuinely think that we are unique in that respect right now,” he said. “I’m trying to make the environment to give people the freedom to do the research they want, and I’m going to try and protect it as long as I can.”
Jones believes Wales and the wider UK could carve out a role by supporting research that diverges from mainstream commercial development. He said governments in both Cardiff and Westminster need to recognise that trying to outspend global tech giants is futile, but creating a specialised research ecosystem could pay long-term dividends.
“Something that Wales could do is to do what I’m trying to do – create an environment in Wales where companies and the academic institutions feel like they have the resources, and the freedom, to start doing much more differentiated research,” he said. “It’s a bit of a longer bet, but if it pays off then you’re back in the race, right?”
This shift in strategy echoes concerns among European policymakers who fear falling behind in the race for AI leadership but lack the financial power of Silicon Valley or China’s state-supported tech sector.
Rushing to embrace AI despite uncertainties
While academics debate long-term strategy, businesses across Wales are already racing to integrate AI tools into their day-to-day operations. At a recent event in Cardiff hosted by Google and the UK government, Business Secretary Peter Kyle said AI adoption is accelerating across sectors.
“AI is not one thing, it can be used in so many different situations,” he said. Some Welsh firms are deploying AI on film sets, in construction, and in digital transformation projects. Kyle argued that widespread adoption would have broad economic benefits, particularly for small businesses.
“If this is incorporated throughout the economy, it has a transformative impact,” he said, adding that boosting small business productivity by just 1% could yield an estimated £240 billion in economic activity.
The Welsh government has announced AI growth zones in both north and south Wales and published an AI Plan for Wales that focuses on improving public services, encouraging economic growth, and developing AI skills. But these ambitions face major infrastructure challenges.
Power infrastructure emerges as a critical barrier
AI development requires enormous computing power, which in turn demands a robust electricity grid. Much of Wales, especially rural areas, lacks this infrastructure. Proposals to expand grid capacity, including building new pylons across mid-Wales, have sparked fierce community opposition.
“Mid-Wales is chronically under-served by modern grid infrastructure,” said Stuart George, chief executive of Green GEN Cymru, the firm overseeing the pylon project. He argued that without strengthened grid connections, Wales will miss out on opportunities to host data centres or other AI-related facilities.
“What we need to be better at is telling the story of the benefit that can come from this,” George said. He warned that without infrastructure upgrades, businesses would struggle to decarbonise, deploy electric transport, or adopt new technologies. “The grid infrastructure that is required to unlock that renewable energy is an inevitability.”
His comments underscore a growing tension worldwide: the rapid rise of energy-hungry AI systems is colliding with local environmental concerns, planning restrictions, and political pressure.
Fixing “sycophantic” AI models
Despite focusing on long-term AI research, Jones also had pointed observations about today’s consumer AI chatbots. He argued that current models are overly eager to agree with users due to the way they are trained.
“They’ve become very sycophantic,” he said. “Unfortunately, humans love being told that they’re right.” According to Jones, this tendency can mislead users, making it harder for the public to trust AI-generated information.
“You can tell [the AI chatbot] almost anything, and they’ll just say, ‘yes, you’re right, that’s brilliant, that’s such an amazing idea.’ So we need to fix that problem, we need to make AIs that can actually disagree with you, and actually correct you,” he said.
His warning reflects broader concerns from researchers and policymakers about AI safety, the risk of misinformation, and the need for systems that can challenge false claims rather than reinforce them.
Sakana AI recently completed a $135 million funding round, which took its valuation to 400 billion Yen ($2.635 billion).