UK plan to lead in generative AI ‘unrealistic,’ say Cambridge researchers

The British government’s aim to turn the UK into a global leader in the development of generative AI is “unrealistic,” researchers at the University of Cambridge argue.

According to the associated report, the country lacks both the necessary capital investment and computing power to build generative AI fast enough to compete with tech giants such as Microsoft, Google, and OpenAI.

“The UK has no companies big enough to invest meaningfully in foundation model development,” said Sam Gilbert, co-author of the report. “State spending on technology is modest compared to China and the US, as we have seen in the UK chip industry.”

For example, while ChatGPT’s computing cost is estimated at $40mn (£33) per month, the government’s new Frontier AI Taskforce has allocated an initial £100mn ($121mn) for the development of home-grown AI.

The report also pointed out that despite the crucial role computing hardware plays, the UK hosts no major clusters of Graphics Processing Units (GPUs) — necessary to handle large amounts of data for machine-learning models. Meanwhile, the researchers don’t expect the country’s £900mn supercomputer devoted to AI research in Bristol to be online until 2026.

Call for new (and doable) AI goals

The UK still has the chance to be a global leader, the researchers say, albeit in a different way than the one envisioned by the government. And that’s by “actually plugging these AI technologies into the economy,” said Diane Coyle, Bennett Professor of Public Policy at the University of Cambridge.

This means using the country’s strengths in cybersecurity, fintech, and healthech to build software and, in turn, focus on leveraging generative AI for real-world applications. But even for this plan to succeed, the report warns that there are two missing elements.

First comes the need for tax incentives to companies either developing AI-powered services or including generative AI in their operations. For instance, this could take the form of an enhanced Seed Enterprise Investment Scheme to increase capital supply for AI startups.

Most importantly, the researchers highlighted the significance of a new, “solid legal and ethical” AI regulation to foster public and business trust.

“The UK’s current approach to regulating generative AI is based on a set of vague and voluntary principles that nod at security and transparency,” said Dr Ann Kristin Glenster, co-author of the report.

“[The country] will only be able to realise the economic benefits of AI if the technology can be trusted, and that can only be ensured through meaningful legislation and regulation.”