Spies in the sky: how worried should we be about the arrival of AI-enabled smart lamp-posts?
A Guardian investigation examines AI-enabled smart lamp-posts in the UK, with experts questioning the commercial viability of the technology's claimed capabilities.
Intelligence analysis by Llama
The Guardian examines Conflow Power Group's iLamp smart streetlights, which the company claims combine solar power, surveillance, and AI datacentre capabilities. Experts call the AI claims hype, while a Nigeria deal raises questions about data control.
Imagine if the street lights could watch everyone and remember every face. Some companies say this is great for catching bad guys, but experts say the fancy computer stuff inside just doesn't work yet — it's mostly hype.
Analysis
The Marketing of Miraculous Lamp-Posts
Conflow Power Group (CPG) markets its iLamp as a triple-threat product: solar-powered streetlight, surveillance hub for vehicle and facial recognition, and "revenue-generating distributed AI datacentre." The first two claims have some technical grounding — automatic number plate recognition has existed for decades, and CCTV is ubiquitous in UK town centres. The third claim, however, sits at the heart of what makes this story an economic concern rather than purely a privacy one.
CPG has reportedly been "extensively covered by worldwide media" but declines further questions, directing journalists to its online material. The opacity is notable. When a company refuses to substantiate technical claims with independent verification, investors and policymakers have reason to be cautious. The reporter's experience at the Warwickshire hospital car park — where neither the local authority nor the hospital will discuss the deployment — illustrates how the rollout has proceeded without meaningful public scrutiny.
When Solar Meets Surveillance Economics
Prof Rabih Bashroush, director of the Centre for Artificial Intelligence and Data Innovation at the University of East London, is direct: "AI datacentres? There's absolutely no chance this will fly." He cites three failures: the technology doesn't work, there isn't enough power, and the economics don't work. Solar generation in Britain is "negligible," he says, fit only for non-critical apps like detecting whether a parking spot is occupied.
Bashroush also flags a deeper structural problem: the difficulty of getting independent expert opinion when most AI specialists work within the industry and have incentives to align with their employers' marketing. A company might make large claims, including AI buzzwords, ahead of a funding round; engineers don't contradict them, local authorities don't understand them, and nobody else is even aware that anything has moved on since CCTV was invented. For investors watching AI hardware, the iLamp story is a textbook case of hype outrunning the underlying physics.
The Data Question Nobody Will Answer
The economic stakes escalate when considering the data layer. CPG has signed an agreement with Nigeria's Katsina state to supply 50,000 iLamps with "additional surveillance potential for both vehicles and faces." CPG chairman Edward Fitzpatrick told the BBC the technology could track missing persons or fight crime, "in partnership with the relevant authority." But who owns the harvested data, where it is stored, and who accesses it remain unanswered questions.
UK Home Secretary Shabana Mahmood has publicly endorsed a "panopticon state" — a state capable of watching citizens at all times. Combined with predictive AI capabilities that could turn reactive surveillance into anticipatory policing, the infrastructure being installed today may outlast the political coalition that authorised it. Big Brother Watch's Jasleen Chaggar warns that local councils are often "sold these tools as magic wands when there are limitations to what they can actually do." For policymakers and capital allocators alike, the iLamp episode is a reminder that AI-inflected infrastructure deals deserve the same scrutiny as any other utility contract — and currently aren't getting it.
Key points
- Conflow Power Group claims its iLamps combine solar power, vehicle and facial recognition, and 'revenue-generating distributed AI datacentre' capabilities
- The company has signed an agreement with Nigeria's Katsina state to supply 50,000 units, but declines media inquiries and directs questions to online material
- Prof Rabih Bashroush of the University of East London calls the AI datacentre claims hype, saying the technology, power budget, and economics do not work in the UK
- UK Home Secretary Shabana Mahmood has publicly endorsed a 'panopticon state', raising questions about who would access the data these lamp-posts harvest
- Local residents in Warwickshire are largely unbothered by the technology, while policy experts warn councils are being sold 'magic wands' with unverifiable capabilities
If the technology matures and is properly regulated, solar-powered lamp-posts with surveillance capability could genuinely help locate missing persons and deter crime, while reducing street lighting energy costs. CPG's Fitzpatrick frames future use explicitly as conditional on partnership "with the relevant authority," suggesting a governance pathway exists.
Without clear data-ownership rules and independent technical verification, iLamps risk installing permanent surveillance infrastructure that outlasts any current political mandate. Experts warn the AI datacentre economics simply do not work, meaning public funds may flow to vaporware while citizens absorb the privacy cost of a system whose capabilities are unverifiable and whose scope can expand silently via software updates.
