UK PM Starmer must find middleground in AI arms race

Keir Starmer outside Downing Street
Image: Shutterstock

In a month that began with Donald Trump announcing an investment worth up to $500bn in US AI infrastructure and ending with China’s DeepSeek wiping $1trn from US tech, UK Prime Minister Keir Starmer now stands at the proverbial AI crossroads.

Along with embarking on a what he calls a “monumental” project alongside OpenAI, Oracle and Softbank, President Trump began his second spell in office by culling 80 executive orders from the previous Biden administration, including the AI Safety Bill.

In doing so, he has removed the guardrails surrounding the development of Generative AI which had previously required developers to share safety test results and consider guidelines to help identify and correct flaws in models, including biases.

On the other side of the Atlantic, Starmer has also set out his vision to “unleash” AI across the UK in a bid to deliver multi-billion pound growth to a stagnant economy desperate for stimulus. 

The PM’s “plan for change” involves AI integrations across the public sector to save on spending and drive growth. This will include the creation of AI hubs, a policy which has been met with concerns over safety and the impact that the resulting massive energy requirements would have on the climate.

This leaves Starmer with a choice. Does he follow the lead of Trump and plough forward with a single minded view of sovereign superiority? Or, does he look closer to home at the UK’s EU neighbours?

In May, the European Council green lit the EU AI Act, which aims to regulate the development of AI in a way that ensures that it is “safe, transparent, traceable, non-discriminatory and environmentally friendly”.

In seeking to become the first regulatory authority to stamp a comprehensive framework on AI regulation, it has been criticsed of focusing on retsrictive parameters that will limit the benefit of business adoption of the technology.

Though Starmer seems committed to a constructive relationship with the EU, it is also a fact that his government is free from being bound to EU legislation after the 2016 Brexit referendum. The PM now now has the chance to tread his own path and set the UK on a journey of AI self-discovery. 

Biases and hallucinations

It goes without saying that there is a significant risk posed by irresponsible AI development. Large Language Models (LLMs), the most common form of AI, rely on vast amounts of data for training. AI hallucination can occur when models fail to grasp the information and therefore generate incorrect outputs.

Recently, Apple was forced to pause its AI notifications as it repeatedly failed to grasp the content of news stories, leading to its AI falsely telling readers that Luigi Mangione, the man accused of killing UnitedHealthcare CEO Brian Thompson, had shot himself.

Similarly, AI systems are at the mercy of developers who decide what information to feed them. Notably, the DeepSeek chatbot, which has shaken Silicon Valley to its core, appears to be censored for topics sensitive to the Chinese government – such as Taiwan and Tiananmen Square.

On a more global scale, it goes without saying that in the wrong hands AI could have a devastating impact for national and international security, which is why DeepSeek is facing scrutiny from the US National Security Council and the US places restrictions on chip exports to certain countries.

Influenced by Silicon Valley czars, the Trump administration views AI as the new economic frontier that must be dominated by the US above all others. Prior to DeepSeek’s revelations, JP Morgan reported that the six principal US tech giants—Meta, Amazon, Tesla, Alphabet, Apple, and Microsoft had spent $500 billion on the development of generative AI investments, with a further $1 trillion reserved for additional investments.

Economic opportunities

On the flip side, the economic benefits of successful AI integration are equally evident. In the space of a few short years, AI has transformed from a buzzword to an integral part of the general public’s day-to-day existence.

Almost every new piece of technology released is powered by AI and tech giants, such as Google, Microsoft and Meta, are increasingly leveraging AI as the main driving force behind their products. 

According to the UK government, estimates from the International Monetary Fund (IMF) suggest that if AI is fully embraced, it could boost productivity by as much as 1.5% per year and, if fully realised, these gains could be worth up to an average of £47bn to the UK each year over a decade. 

The UK’s AI masterplan has also been backed by £14bn in committed investment from Vantage Data Centres, Nscale and Kyndryl. In the US, tech giants OpenAI and Oracle, as well as investment fund SoftBank, have pledged $100m as part of a new joint venture, however, this could rise to five times this amount.

Despite the revelations of DeepSeek’s cost-cutting AI models- which it claims were deveIoped for as little as $6m – it is clear that investors have money to burn when it comes to AI and the UK must position itself at the front of the queue for future capital as it looks to remain at the table of the global elite.

On the other side of the world, India has also positioned itself as a key overseas market for tech giants such as Microsoft, which has committed to investing $3bn to expand its AI and cloud services in the nation, as well as train an additional 10 million people in AI.

Finding the middle ground

Although it may seem like fence-sitting, there is a happy medium to be found between the two approaches. Although there are genuine concerns regarding biases and hallucination, the economic potential of AI in the public sector is undeniable.

Growth in the UK appears to have almost ground to a resounding halt and as the world clamours for AI, it would be economic suicide for Starmer not to strive for the UK to be at the centre of the AI revolution. 

While AI looks to be the next frontier for the ongoing arms race between the US and China, there is room for the UK to slide its way on to the podium and make itself the go to location for AI investment outside of the two tech heavyweights.

Speaking to the BBC, Science and Technology Secretary, Peter Kyle, rightfully states that there is no reason why the UK cannot create tech companies on the scale of giants such as Google, Amazon and Apple.

He also promised that the UK will be “agents of change” rather than “cautious bystanders”.

As it stands, the UK government has adopted a cross-sector and outcome-based framework for regulating AI that is implemented in each sector by applying laws and issuing supplementary regulatory guidance. 

This adaptability must remain as the UK strikes a balance between innovation and ethical oversight if it is to reach its goal of becoming an AI powerhouse. Starmer must understand the threat posed by AI but equally his, and the nation’s, ambition must not be quelled by AI’s most vocal critics.