Conflow Power Group has installed solar-powered iLamp streetlights in a UK hospital car park that can host cameras, sensors, and edge AI computing—but neither the hospital nor the public can confirm what the poles are actually watching. The company, which also plans a 50,000-unit deployment in Nigeria, markets the iLamp as a ‘distributed AI datacenter,’ raising urgent questions about hidden surveillance capabilities that could be activated without warning.

Putting a Smart Pole in Every Parking Lot

The iLamp is, on its surface, an environmentally friendly upgrade: a solar-powered streetlight that can operate off-grid. But CPG’s own literature describes it as a platform that can “double as a revenue-generating distributed AI datacenter” and support smart-city applications such as traffic management, public-safety sensing, and, where regulations permit, AI-enabled cameras.

The first UK installation is at a Warwickshire hospital car park—an unassuming testbed that has largely flown under the radar. When The Guardian’s reporter visited in late July 2026, drizzling rain and indifferent locals greeted them. “The town centre is already full of cameras, and it’s fine. You don’t mind them being there unless you’re up to no good,” said Karl, a 52-year-old pub-goer. A radiologist waiting at the bus stop expressed unease about AI’s broader trajectory but accepted that “we have to move on and accept that things change.”

Meanwhile, CPG has agreed to supply 50,000 iLamps to Katsina state in Nigeria. Chairman Edward Fitzpatrick told the BBC that the poles could eventually track missing persons or fight crime—but only “with the relevant authority.” The gulf between a hospital lamp-post and a state surveillance network is narrowing by the pallet.

The Surveillance Smorgasbord: What’s Already Under the Hood

A lamp-post can be more than a light fixture. The iLamp’s housing has room for:

  • High-definition cameras
  • Automatic number plate recognition (ANPR) sensors
  • Facial recognition modules
  • Behavioral analytics software (gait analysis, aggression detection)
  • Environmental and audio sensors

Critically, these are not all active in every installation—and that’s the point. The hardware can sit dormant, humming along as a simple light, until someone with authorization decides to flip a remote switch. “A lot of these technologies and infrastructures have capacity to add function later,” Jasleen Chaggar, senior legal and policy officer at Big Brother Watch, told The Guardian. “You can just turn the switch and it’s all there overnight.”

That distinction separates traditional CCTV from live facial recognition. CCTV records footage for later review; live facial recognition compares every passing face against a watchlist in real time, treating your identity like a fingerprint. Embed that capability in a lamp-post, and you’ve created an ambient web that can follow you from the hospital to the high street without ever installing a new camera.

Why You Should Care Even If You’re ‘Not Doing Anything Wrong’

For the millions who visit UK hospitals each year, the presence of an iLamp raises questions that go unanswered because no privacy impact assessment has been made public. A camera pointed at a hospital entrance or car park doesn’t just record vehicles; it records who is seeking medical care, who is visiting loved ones, and what patterns emerge over time. Under UK data protection law, health-related inferences can be especially sensitive.

The indifference of Karl and others in the car park is understandable—we’ve all grown accustomed to CCTV. But facial recognition operates on a different plane. “It’s much more like a fingerprint or DNA,” Chaggar said. “It’s not like a passport—you can’t get another face.” Misidentification in retail has already led to innocent people being blacklisted from stores without explanation. Imagine that logic applied to a hospital: a faulty algorithm flags you as aggressive, and your next visit triggers a security intervention.

For IT professionals managing campus networks, the iLamp is also an IoT device with an operating system, a management plane, and a remote-access backdoor that your team may not control. If a pole is compromised, it could become a pivot point into a hospital’s broader network—a risk that feels chillingly close to the Mirai botnet attacks that weaponized connected cameras years ago.

The Tech Reality Check: Can a Lampost Really Be an AI Datacenter?

CPG’s claim that the iLamp can function as a distributed AI datacenter has drawn flat rejections from technical experts. “There’s absolutely no chance this will fly: the technology doesn’t work, there’s not enough power, the economics won’t work,” Professor Rabih Bashroush, director of the Centre for Artificial Intelligence and Data Innovation at the University of East London, told The Guardian. In Britain’s dreary climate, the solar energy generated is “negligible.”

That doesn’t mean edge computing is impossible. A pole could run narrow, non-mission-critical applications like parking-spot occupancy detection or basic environmental telemetry. But facial recognition that requires continuous, high-resolution video processing and real-time database lookups? “Forget it—this has no legs,” Bashroush said.

Still, the gap between marketing and reality can be exploited. A local authority might buy iLamps for lighting, seduced by the idea of future AI revenue, only to find the hardware supports surveillance capabilities that exceed its original remit. For procurement teams, this means digging into the technical specifications—not the brochure—and asking hard questions:

  • What is the compute hardware? CPU/GPU model, wattage, and thermal constraints?
  • What operating system runs on the pole, and who patches it?
  • How does remote access work? Is there a VPN, and who holds the keys?
  • What data leaves the pole? Where is it stored, and for how long?
  • Can feature activation be tied to a public notice and privacy assessment, not a silent firmware update?

Without these answers, a lamp-post is a black box with a power cord.

How We Got Here: From CCTV to Ambient AI

Britain’s relationship with public surveillance is decades old. The UK has one of the highest densities of CCTV cameras in the world, normalized through crime-prevention narratives. Automatic number plate recognition arrived later, feeding the National ANPR Data Centre. Police forces began trialing live facial recognition around 2016, often without explicit legal frameworks, prompting legal challenges by civil liberties groups.

In 2025, Home Secretary Shabana Mahmood said she believed in a “panopticon state”—a state that could watch you at all times—arguing that “you can’t enjoy any of your liberties in any country if you’re not safe.” That philosophy, applied to infrastructure procurement, encourages exactly the kind of “slippery pole” the iLamp represents: sell a light, install a camera node, and let the legal justifications catch up later.

Hammersmith and Fulham Council in London recently upgraded its CCTV network to allow embedding of facial recognition, suspicious shopping behavior detection, and aggression monitors. Residents noticed the new beetle-like cameras whirring as they walked past—a visible change. The iLamp removes even that visibility. The shell stays the same; the capabilities change inside.

What You Can Do Right Now

If you’re a resident, patient, or visitor:

  • Ask your local council or NHS trust whether iLamps or similar smart poles have been installed, what sensors are active, and whether a data protection impact assessment (DPIA) has been published.
  • Submit a Freedom of Information request to uncover contractual details—you don’t need to be a lawyer to ask.
  • Raise the issue with elected representatives—councillors, MPs, and hospital governors can press for transparency.
  • Be vocal in public consultations about smart-city initiatives; procurement decisions made today lock in the surveillance architecture of tomorrow.

If you’re in IT or procurement:

  • Include enforceable technical limits in contracts. Specify that no biometric or behavioral analysis feature may be activated without a formal change-control process, a new DPIA, and public notice.
  • Demand an asset inventory for every pole that includes hardware spec, OS, patch cadence, and remote-access methods.
  • Insist on data-ownership clauses. The tragedy of the Netherlands garbage-collection data, where the supplier walked away with the data when the contract ended, is a warning.
  • Plan for sunset. If you terminate the contract, how do you retrieve your data? What assurances exist that the poles won’t keep phoning home?

For the technically curious, if you spot an iLamp, note the model number, visible ports, and any network equipment. Sharing that information with watchdog organizations can help map the spread of surveillance-ready infrastructure.

The Growing Net of Smart Surveillance

The iLamp is one product, but the blueprint is spreading. Smart poles are popping up in smart-city pilots from Barcelona to Singapore, each with its own menu of sensors. The UK’s test case—a hospital car park—is especially unsettling because of the sensitive location. If the privacy debate doesn’t happen now, we may find ourselves living under an invisible mesh of eyes that we never agreed to, and cannot easily unplug.

Watch for how the Information Commissioner’s Office responds to queries about these installations, whether Parliament’s Science and Technology Committee takes an interest, and how quickly other local authorities adopt similar platforms. The question isn’t whether a lamp-post can spy on you today. It’s whether, after the hardware is in place, anyone will be brave enough to leave the switch off.