London's urgent race to make AI safer

A wide image of the front of King's Cross station

Long before artificial intelligence became a regular fixture of news headlines, London had quietly been shaping itself as a centre for AI safety.

As AI systems become increasingly powerful, a growing ecosystem of researchers, policymakers and campaigners in the capital is focused on a related challenge: ensuring these models benefit society rather than cause harm.

Increasingly described as Europe's leading AI hub, London is home to Google DeepMind, while Anthropic has also established a major base in the capital.

According to the London AI Hub, a co-working space and community for AI entrepreneurs, up to 50 organisations in the capital now focus on AI safety.

They include the UK's AI Security Institute (AISI), described as the first state-backed organisation for assessing leading models before and after release for risks like cyber misuse, and Apollo Research, which looks at the potential for AI models to act against human interest by lying or cheating.

The London Initiative for Safe AI (LISA), a non-profit organisation that supports and connects people working in AI safety, is also set to relocate to a large headquarters in King's Cross, part of the capital's growing "AI corridor".

But as warnings over the potential dangers of the technology ramp up, will these efforts genuinely make a difference on the global stage?

A man stands in a lab looking at the camera

The government has announced fresh investment worth hundreds of millions of pounds in AI safety and security, including work on threats ranging from cyber-attacks to the possibility of advanced systems acting beyond human control.

One concern increasingly occupying researchers is in the field of biology. While AI could help accelerate the discovery of new medicines and vaccines, experts have also warned it could make it easier to develop harmful pathogens.

As part of its response, the government is to launch an AI Biosecurity Lab. One of the organisations involved is Basecamp Research, an AI biology company based in a former watch factory in Farringdon, central London.

Working alongside OpenAI and Anthropic, Basecamp Research will help develop antibiotics and vaccines to protect against future pandemics and get "ahead of emerging threats".

Co-founder Oliver Vince believes it will "put the UK at the forefront of the intersection of AI and biology", particularly in defending against harmful pathogens.

He says that London is a key player in AI security due to its "opportunity, culture, history and expertise".

A woman in a lab coat does work in a lab

The firm is one of many AI startups clustered in the capital.

Rupert Elston from the London AI Hub, a community and co-working group, estimates up to 50 of these focus on AI safety, meaning the city has become "not just a player in the global AI game, but also something of a referee".

"The boundaries are very clearly defined. That creates an environment where people want to come and build, and helps attract talent as well."

'London makes me optimistic'

Prof Hannah Fry, a woman with long red hair. She is wearing a dark blue sweater and is sitting behind a BBC Desert Island Discs-branded microphone

The capital's growing role in AI safety, and using the technology responsibly, is attracting increasing attention.

Speaking on BBC Radio 4's Desert Island Discs, the mathematician, author and science broadcaster Prof Hannah Fry said she remained optimistic about AI because "something is happening in London".

"The UK punches well above its weight in scientific expertise and technological ability. We are a global hub."

She added that London was on a different path from San Francisco's "race-to-the-bottom hyper-capitalism", instead focusing on "ethical technology for good, focused on things people care about – from finding cures for cancer to developing better batteries. And all of that is happening right here on our doorstep."

Fry added: "I think we'll be buffeted by some really big forces over the next few years, but I know these scientists and I'm excited by what they're coming up with."

The infrastructure challenge

But London's AI safety ambitions face a number of obstacles.

Gina Neff, professor of responsible AI at Queen Mary University of London, says one of the biggest challenges for researchers is gaining access to cutting-edge AI models before they are released to the public.

"It shouldn't just be a few people in northern California who set the terms of the future for the rest of us," she says.

She points to the UK's track record of developing regulatory "sandboxes" that allow researchers and policymakers to test emerging technologies in controlled environments before they are deployed more widely.

However, Neff, who is also deputy chief executive of Responsible AI UK, warns that building a world-leading AI ecosystem will require the data centres to support it.

And these centres come with their own concerns, particularly the vast amounts of energy they consume.

A report commissioned by the Greater London Authority, London's Data Centre Future Growth Forecasts and Strategic Insights, identifies significant pressures on the capital's energy network, warning that "energy is the single most critical constraint on London's data centre growth".

A large grey warehouse

Despite the challenges, many of London's leading universities are also keen to keep pace - and responsibly.

At Imperial College London, president Hugh Brady says "AI is rapidly permeating every aspect of what we do. It's embedded in our education programmes, driving and accelerating much of our research, and at the heart of many of our start-ups."

Brady estimates that more than 750 people at Imperial are working either directly on large language models (LLMs) or on their applications.

'The train has left the station'

But while he recognises AI's enormous potential, he is also clear about the risks.

"In the wrong hands, AI could be used to design something harmful to humans, the food chain, the environment or our water supply... something we haven't encountered before."

Even so, he believes AI must be part of the solution.

"The train has left the station," he says. "All of this has to happen at greater speed than ever before."

Five men hold up a banner that reads "PAUSE AI"

Activist Joseph Miller argues that London's historical links to effective altruism – a movement to find the most impactful ways to help others – alongside its leading universities and big student communities, have helped make it one of the world's leading centres for AI safety.

He first staged a protest outside Google DeepMind's offices with a friend in 2021.

Five years later, he runs Pause AI UK, a group calling for a halt to the development of frontier AI models until, as he puts it, "we know how to make them safe".

The movement has grown rapidly since then. "We have hundreds of people joining every month," he says.

Pause AI's last protest in London, held in February, attracted about 300 people. Their next demonstration, planned for Saturday, is expected to draw a larger crowd.

A spokesperson for the Mayor of London Lord Khan said: "London has the talent, universities and businesses to lead the way in using AI responsibly.

"The government's additional £120m for the AI Security Institute will strengthen the UK's ability to test advanced AI systems and understand the risks they pose.

"The mayor wants Londoners to share in the opportunities AI brings. That's why he set up the London AI and Jobs Taskforce to look at how AI is changing the way we work and how Londoners can prepare for those changes."

Original source London's urgent race to make AI safer

Back to home