Researchers created a $7 device capable of locating hidden cameras within seconds.

      It operates by sweeping light rather than merely shining a spotlight on reflections with a 94% accuracy rate.

      Hidden cameras are frequently discovered in discreet locations, concealed in pens, clocks, chargers, or picture frames within hotel rooms and rentals. This prompted researchers at KAIST to create a solution that is more affordable than a good lunch. Their innovative device, known as SweepLED, transforms any smartphone into an effective hidden camera detector using an attachable LED case that costs less than $7 to produce.

      Why your existing hidden camera detector might not be effective

      Most handheld detectors utilize a straightforward method: shine a light on an object and look for bright reflections bouncing back. The issue is that glass, metal, and shiny plastic also reflect light, leaving you to guess if that gleam is a camera lens or merely a chance occurrence while squinting at chargers and clocks.

      SweepLED removes all uncertainty. Instead of moving both the light and the camera angle simultaneously, it keeps the phone's camera stationary and sweeps the LED light from various angles. This is vital because a camera lens features an internal structure, aperture, sensor, and layered glass, causing its reflection to warp and deform in a specific manner as the angle of light changes. Common shiny surfaces do not exhibit this behavior.

      How SweepLED operates

      SweepLED captures the entire light sweep as a brief video, which it then processes through an AI model designed to identify the distinctive lens distortions, pinpointing precisely where a hidden camera might be located. Researchers tested it against 30 ordinary items typically found in a hotel room or rental, achieving approximately 94% accuracy in detecting hidden cameras, all within five seconds per object.

      A recent study from UCL revealed that users of consumer-grade hidden camera detectors still overlooked 59% of devices during testing, making SweepLED’s 94% accuracy in the lab a promising indicator, with real-world testing anticipated soon.

      Led by Professor Jun Han at KAIST’s School of Computing and collaborating with researchers from the National University of Singapore and Singapore Management University, the project was showcased earlier this year at ACM MobiSys 2026. Although it is not yet a consumer product, the concept is straightforward enough that it could feasibly appear in a real device in the near future.

      Manisha Priyadarshini is a tech and entertainment writer with over nine years of editorial experience.

      Why the Next Generation of Home AI Features Wheels and Emotion

      OlloBot is advancing technology from the countertop into everyday life with its expressive new companion. For the past ten years, interacting with home artificial intelligence has adhered to a predictable, rigid framework: a person invokes a wake word, requests the weather or to switch off the lights, the speaker responds, and the interaction concludes. While this system is functional, it keeps smart devices strictly transactional, relegating them to the role of tools activated for specific tasks instead of systems that seamlessly integrate into a living environment.

      OlloBot is pursuing a fundamentally different approach with the OlloNi SS1, an embodied AI companion, which refers to artificial intelligence embedded in a physical, mobile form rather than confined to a screen.

      AI aims to convert simple phone videos into cycling performance data that serious riders typically pay thousands for

      Cycling can be an expensive pastime. However, as a cyclist becomes more serious, the costs can escalate significantly for measuring performance. In professional contexts, equipment such as instrumented pedals, bike computers, and heart-rate monitors become essential, all aimed at extracting more data from each ride. Now, researchers at La Trobe University (via TechXplore) are investigating whether a single piece of equipment that most individuals already possess could potentially manage a surprisingly complex aspect of that task. This isn't specialized equipment; it's a small device that fits in your pocket—your smartphone. Researchers at the Holsworth Biomedical Research Centre are creating AI models capable of estimating the forces exerted by a cyclist on the pedals through video motion capture. Cyclists could simply record themselves with a smartphone and receive biomechanical insights that presently require much more specialized apparatus.

      UK health advocates caution that AI transcription tools may put patients in jeopardy

      AI scribes are increasingly being adopted in medical offices, promising to reduce boring paperwork and allocate more time for patient care. However, mounting evidence indicates that these tools can introduce new risks for both patients and healthcare providers. AI scribes are distorting diagnoses and prescriptions.