Australia’s AI surveillance debate puts icetana in the frame

Australia's AI surveillance debate has intensified after the Administrative Review Tribunal overturned a 2024 Privacy Commissioner ruling that Bunnings breached privacy laws by using facial recognition, and Coles and Woolworths confirmed testing of the technology. Perth-based AI company icetana (ASX:ICE) sees opportunity in its software that monitors existing CCTV for unusual events rather than identifying individuals, addressing the problem that most cameras are not being monitored, according to CEO Kevin Brown.

Australia’s AI surveillance debate puts icetana in the frame Bunnings, Coles and Woolies eye smarter cameras Privacy fight puts facial recognition on notice AI surveillance opens opportunity for icetana Australia has a CCTV problem. Not because there aren’t enough cameras. There are plenty of those. The problem is what happens after they are installed. A shopping centre, university or other large facility can have hundreds or even thousands of cameras feeding footage into a control room. But there are only so many screens a security officer can watch at once, and only so many hours a human can stare at mostly uneventful footage before something important slips past. Artificial intelligence offers an obvious solution: let the software watch the cameras and tell humans when something happens that is worth viewing. But that solution is now fuelling a much bigger debate in Australia over just how smart – and intrusive – those cameras should become. What exactly should the AI be looking for? Facial recognition faces a reckoning Bunnings has become Australia’s biggest test case for one answer. The hardware giant used facial recognition technology to scan shoppers and compare them against an enrolled list of people it said had been involved in theft, fraud, threats or other incidents. Australia’s Privacy Commissioner ruled in 2024 that Bunnings had breached privacy laws. But in February this year, the Administrative Review Tribunal overturned that finding – deciding the retailer was entitled to use facial recognition for the limited purpose of combating serious retail crime and protecting staff and customers. The tribunal did, however, find shortcomings around how customers were notified and how privacy information was handled. That ruling suddenly made the issue much bigger than Bunnings. This month, Coles and Woolworths confirmed they had conducted testing of facial recognition technology https://www.abc.net.au/news/2026-08-17/coles-and-woolworths-to-test-facial-recognition-technology/107045036 as retailers look for new ways to deal with aggression, violence and theft. Neither has committed to rolling it out across Australian stores. But the fact Australia’s two supermarket giants are even examining the technology shows how quickly AI-powered surveillance is moving into the mainstream. And retail is only part of it. WA Police has been trialling live AI facial recognition in public, prompting concerns from privacy and legal experts about mass scanning, false matches and safeguards around the technology. The debate, in other words, is now well and truly alive. Smarter cameras don’t have to mean Big Brother There is, however, an important distinction that can get lost whenever “AI surveillance” is thrown into one bucket. AI does not necessarily need to answer: Who is that person? It can instead ask: What is happening? That distinction sits at the heart of the opportunity for icetana ASX:ICE https://stockhead.com.au/company/icetana-ice/ . The Perth-based AI company develops software that works with existing surveillance systems. Its technology learns what normal activity looks like across individual cameras and brings unusual events to the attention of security operators. In other words, it helps security teams find the moments that matter without having to watch every camera themselves. Chief executive Kevin Brown says the underlying problem is brutally simple: “Most cameras are not being monitored.” Australian control rooms can have hundreds of cameras but only a fraction visible to operators at any one time. Put 50 live feeds in front of someone and concentration quickly becomes the problem. Multiply that into hundreds or thousands and the maths stops working. icetana’s software effectively filters that fire hose. Look for the event, not necessarily the identity That distinction becomes particularly interesting as businesses try to balance safety with increasingly vocal privacy concerns. A facial recognition system may try to determine whether somebody entering a store matches a known person. icetana’s technology can instead look for activity that deserves attention. Its platform has been developed to detect unusual behaviour and events, while newer capabilities include aggressive behaviour, floor hazards and indicators associated with shoplifting. Brown describes icetana’s approach as generalised and self-learning. “You switch it on, and 24 hours later, it just works.” The company says its original algorithms have been developed using more than 700 million hours of CCTV footage, giving it a sizeable body of real-world video experience accumulated over more than a decade. Its software is also camera agnostic, meaning customers don’t need to rip out their existing CCTV estate to use it. That’s a useful advantage in a market where much of the camera infrastructure is already installed. Commercial case starts to take shape For investors, this story would be less interesting if smarter surveillance remained just a clever technology looking for customers. The numbers suggest icetana is beginning to turn that technology into a more scalable business. Its software is deployed across more than 19,000 cameras at over 70 sites in more than 15 countries. Annual recurring revenue reached $2.6 million in the June quarter, up 33% year on year, while quarterly revenue rose 37% to $648,000. The company also landed the first commercial orders for its new Antara Core platform, with Australian partner Millennium Services Group placing two five-year orders worth a combined $617,000. Read more about that here: icetana AI turns smarter surveillance into global sales Antara Core is particularly relevant to the privacy conversation because it can run AI on-premise. For organisations handling sensitive footage, that means the video processing can remain inside their own infrastructure rather than sending footage elsewhere. As Brown puts it: “Most of our customers do not want a single pixel leaving the building.” That could become an increasingly valuable selling point as organisations want the benefits of AI without creating another headache around where surveillance footage is being processed. From security alerts to actual action There is another layer to the icetana story. AI watching cameras is useful. AI triggering a response could be much more useful. The company’s floor-hazard detection provides a good example. A camera can detect a spill or obstruction. Instead of relying on someone to notice it on a screen, the system can generate an alert. Through work with SoftBank Robotics, icetana is also exploring workflows where an alert could ultimately send a robot to deal with the problem. That shifts the proposition from surveillance towards automated facilities management. Brown says this is where the potential market begins to expand beyond simply counting cameras. And icetana is trying to reach that market increasingly through partners rather than building a huge global sales force itself. SoftBank Robotics has expanded its distribution relationship with icetana into North America – adding to agreements covering Japan, Australia and New Zealand, wider Asia-Pacific, and the UAE and Saudi Arabia. Japan offers an early indication of what that model could deliver, with annual recurring revenue through SoftBank Robotics reaching about $680,000 in its first year as exclusive distributor. icetana isn’t alone in the smart-camera race icetana isn’t the only ASX small cap making cameras smarter. SenSen Networks ASX:SNS https://stockhead.com.au/company/sensen-networks-sns/ uses AI and computer vision to analyse camera feeds, including detecting people, vehicles, objects and unusual activity. Its technology is used across areas including smart cities, transport, retail and security. Spectur ASX:SP3 https://stockhead.com.au/company/spectur-sp3/ combines AI-enabled cameras and software with autonomous surveillance systems, particularly for remote and off-grid sites. Its AI can identify people and vehicles, filter false alarms and trigger automated alerts or deterrents such as lights and voice warnings. Neither is a direct like-for-like competitor, but together they point to the same trend: CCTV is shifting from simply recording footage to AI that can identify what’s happening. At Stockhead we tell it like it is. While icetana is a Stockhead advertiser, it did not sponsor this article. This story does not constitute financial product advice. You should consider obtaining independent advice before making any financial decision. Related Topics UNLOCK INSIGHTS Discover the untold stories of emerging ASX stocks. Daily news and expert analysis, it's free to subscribe. By proceeding, you confirm you understand that we handle personal information in accordance with our Privacy Policy https://stockhead.com.au/privacy-policy/ .