Lessons from London Tech Week: The next AI question is who gives it permission to act

By Isabelle Dann, Associate Director, Aspectus Group 

London Tech Week brought crowded stages, countless sovereign AI commitments, major infrastructure announcements, and enough optimism to fill Olympia several times over. More than £6bn of new investment and around 8,000 jobs were announced during the week, alongside a £1.1bn AI hardware plan designed to strengthen the UK’s chip and compute capacity.  

Across sessions, it became clear that the next phase of AI will be shaped by permission as much as performance. A key question, then, is what allows technology to move from promise to practice? 

Lovable co-founder Fabian Hedin described AI as a moment when “the person closest to the problem can now solve it”. The observation stayed with me because it captured both the opportunity and the challenge. Technology can make action easier, but it does not remove the need for trust or accountability.  

This thread appeared in almost every session I attended. Many of these challenges have a communications dimension. Organizations increasingly need to explain how decisions are made and who remains accountable when AI is introduced. 

Permission in practice 

In public services, it was about public consent. In deep tech, it was companies building serious capability faster than customers and institutions can absorb. In healthcare and neurotechnology, it was whether extraordinary capabilities can earn trust through evidence and patient benefit. 

Adoption suggests a relatively clean journey from pilot to deployment – but permission is more complex. It asks who needs to trust the technology, who carries the risk, who has the authority to say yes, and who sees the benefit clearly enough to accept change. 

Take public sector AI. Several London Tech Week sessions returned to a simple question: why should the quality of service people receive from the state feel so much worse than what they experience from the private sector? AI has obvious potential here – from housing sensors that identify damp and mould, to tools that help police or local authorities allocate scarce resources. Still, the public case for AI will not be won through abstract efficiency claims. It becomes stronger when attached to visible service improvements: safer homes, faster answers, more responsive institutions. 

Conversations about AI are often strongest when they begin with the service challenge being addressed, alongside a clear explanation of how the technology contributes to a better outcome. People are more likely to accept AI when they can see the problem it solves and the safeguards around its use. 

Sovereignty beyond the slogan 

Discussions around sovereignty were revealing, partly because the word itself kept stretching. Realistically, no country can own every layer of the AI stack – and the UK should be wary of turning sovereignty into a slogan. It’s more important to understand where dependence becomes a risk. Is it compute? Energy? Chips? Cloud infrastructure? Public datasets? Technical skills? Scale-up capital? Procurement? Once the dependency is named, the policy choice becomes clearer. Until then, sovereignty risks becoming a comforting word for an uncomfortable set of trade-offs. 

A government cannot credibly ask citizens, companies, or public bodies to embrace AI if it cannot explain the dependencies behind it. Permission depends on clarity about where control exists, where exposure sits, and which compromises are being made. 

When systems slow adoption 

Various London Tech Week discussions highlighted the friction between technological progress and the people expected to adopt it. Three-to-six-year procurement cycles make little sense in a software environment where products improve through iteration. Small companies struggle to bridge the gap between pilot and production, as their technology depends on testing, feedback, and controlled use. Investors become impatient while buyers wait for certainty that may never arrive. 

Many emerging technology companies face the same structural problem: the person who understands the need may not hold the budget; the buyer may not understand the technology; the procurement process may not fit the pace of development. Permission to deploy is distributed across a long-standing system that often prioritizes certainty over innovation. The result is that promising technologies often move more slowly than the problems they were built to solve. 

Technology must earn its place 

The health and neurotechnology sessions offered the most human version of this argument. In discussions about Huntington’s disease, brain-computer interfaces, and functional neurosurgery, the promise was far removed from productivity dashboards. It was about slowing the progression of a devastating disease, restoring speech, giving someone the ability to control a wheelchair, or returning a degree of agency to patients who have already lost too much. 

These are the stories that make technological progress feel meaningful. They also show why permission matters. When a device enters the brain, or an AI system shapes clinical decision-making, capability is only one part of the story. The public expects evidence and patients need dignity; clinicians require confidence and regulators must have a clear account of risk. In short, technology must earn its place. 

The strongest technology stories are rarely about capability in isolation. They explain the conditions that allow the capability to matter. What problem is being solved? Who needs to trust the answer? What changes in the real world if it works? What must be true for someone to say yes? 

That is where communications can help. In a crowded AI conversation, the challenge is to explain why a technology deserves permission to be adopted. London Tech Week showed plenty of ambition; the next test is whether that ambition can earn consent. 

Want to continue the conversation? Read our CIO Audience Insights series for a deeper look at how senior technology leaders assess AI and other emerging technologies. If you’d like to discuss your own AI communications strategy, get in touch. 

Key takeaways 

Why is permission becoming the biggest challenge for AI? 

Because organizations must earn trust from customers, citizens, regulators and employees before AI can be widely adopted.  

What helps people trust AI?

Clear communication about the problem being solved, the safeguards in place, accountability and the real-world benefits delivered.

Why is AI sovereignty more complicated than it sounds?

True sovereignty is understanding critical dependencies – such as compute, cloud infrastructure, skills and data – and making informed policy choices around them.

How can communications accelerate AI adoption?

By explaining not just what AI can do, but why it deserves trust, who benefits and how risks are managed.

About the author

Izzy, based in our London office, began her career as a journalist and now works with tech founders, C-suite leaders, developers, and investors to shape their narratives and build their brands. She brings together sharp storytelling, strategic insight, and strong media relationships to deliver creative, high-impact communications. 

In addition to earned media, Izzy develops in-depth reports and thought leadership, with a focus on telecoms, deep tech – including artificial intelligence (AI) and cybersecurity – and venture capital. 

You can find her on LinkedIn here and can be reached at isabelle.dann@aspectusgroup.com.