It began with a threat, then descended into squabbles about blame. This year, London Tech Week was so riven with conflict, it made Love Island feel like a yoga retreat. The event kicked off with Sir Keir Starmer picking a scrap with Google and Apple in his keynote, when he insisted the titans of technology block access to naked images on smartphones for under-18s. He set a deadline of three months to activate these hidden settings, which are already in most phones, or face new legislation. Everyone knew this was merely a sighter for the real whammy, which duly arrived a few days later, when the Government confirmed it will ban social-media apps — including YouTube — for children by next spring.
Things got even livelier once the idea of sovereign AI reared its head. Put simply, this describes the extent to which a country can deploy AI without becoming overly reliant on other nations. The problem is that people don’t agree on what sovereign control involves. The Government talked about investing in AI skills and infrastructure, which largely seems to mean buying semiconductor chips from foreign firms. Other panel experts queried if we have enough energy or computer power. While it’s unrealistic for UK firms to compete with OpenAI, there were widespread concerns about allowing the big beasts of America to dominate AI, as they do with cloud computing or social media. As one speaker put it, today’s friends might not be tomorrow’s allies. And if you think that sounds unlikely, last week the Trump administration issued an export-control directive that forced Anthropic to restrict access to the newest model in its Claude software to the US only.
It was good to see a UK start-up, Cosine, unveil Lumen Sovereign, an AI model that runs entirely on British infrastructure
With AI getting baked ever more deeply into everything we do, these stakes will only grow higher. For many organisations, the question rapidly becomes: “Who owns the operational intelligence layer?” warned Matt Harris, Managing Director of HPE. To counterbalance this, it was good to see a UK start-up, Cosine, unveil Lumen Sovereign, an AI model that runs entirely on British infrastructure such
as Isambard, which is one of Europe’s top supercomputers. It has a coalition of blue-chip backers, including BAE Systems and BT. Rather than compete with ChatGPT, this is a secure platform for sectors like banking or defence.
Stage all set for AI
Make no mistake, the roll-out of AI is now considered such big potatoes at London Tech Week that the main stage was this year rebranded as the AI Stage. And these rollouts are beginning to get serious, even if you’ve not fully noticed them. For instance, London unicorn Eleven Labs confirmed its next-gen AI “voice agents” already serve four million customers in 30 languages. If you talk with Deliveroo, Revolut or Trainline — all of which are clients — about, say, a refund, it’s plausible you’ll already be speaking to a sophisticated bot. At London Tech Week, the company announced a deal with the Department for Science, Innovation and Technology (DSIT) “to find new ways to use voice AI to improve public services”. So, next time you suspect you might not be conversing with a human when you call HMRC about a tax return, there’s a fair chance that you are not.
The Crown courts are also set to feel the full force of AI. In another keynote, David Lammy, the Deputy Prime Minister and Justice Secretary, outlined the emergence of AI legal assistants and tools for judges. He said the goal is to “augment” admin and to support judges to determine the “preparedness” of cases for trial. Tell that to the fans of Maccabi Tel Aviv, whose travel plans to a football match in Birmingham were thwarted by a UK police force on the basis of iffy AI evidence last year. Still, if this reduces backlogs, it’ll prove popular and be no more controversial than trial without jury.
Princely intervention
The elephant in the room with all these government pronouncements is whether Starmer, Lammy or indeed anyone else will still be in post by the time their actions crank into action. One person whose role is not at risk is Prince William, who put in a surprise appearance at the show. He launched the Homelessness Data Lab, an initiative to prevent people losing their homes. The premise is simple. Many of the reasons why people become homeless can apparently be predicted by analysing data to identify the warning signs before their situation reaches crisis point. The Data Lab, which is in partnership with SalesForce and LandAid, brings together 25 private and public sector organisations to connect these dots and then intervene in a variety of ethical ways.
The Prince told the audience: “Many of your customers, your clients, will be using data through banking apps, through their phones. I’m not sure you realise how much that data can be used to predict and see problems with potential homelessness before they actually arise.” This makes a change from scouring the data for commercial purposes and is reminiscent of the concept of “future crime” in Minority Report, the Tom Cruise film.
With so much of this year’s event consumed by angst, let’s finish on a positive note. Among the many inspirational medtech stories at London Tech Week, it was fascinating to hear directly from Max Jaderberg, President of Isomorphic Labs, which is a spin-off from Google Deepmind. This London-based firm is best known for its AI model, AlphaFold, which predicts protein structures and enabled its CEO, Sir Demis Hassabis, to win the 2024 Nobel Prize in Chemistry. Isomophic now makes next-gen AI medicines, with the noble goal of eradicating all disease — and Jaderberg certainly sounded bullish about the prospects of achieving this incredible feat soon. In these uncertain times, let’s cling onto that.
