A viral technical investigation is forcing LG to defend the roughly 49 million webOS-powered smart TVs it has sold in the United States, after researchers alleged the sets quietly capture audio, generate text transcripts, and scan home networks even when the television appears to be off.
What the researchers found
The report, a 135-minute deep dive published by the YouTube channel Gamers Nexus in partnership with Level1Techs and independent security researchers, examined LG OLED televisions running webOS and concluded the devices continued capturing microphone audio after their voice-recognition feature was invoked, including during standby mode when the screen is dark. The team said it recovered plaintext transcripts derived from audio recorded by the TV, in some cases persisting 10 to 15 seconds after silence following the “Hi LG” wake phrase. One researcher described finding audio queued locally for later upload, arguing the pattern suggested the data was being retained for “logging, analytics or model training” rather than simply executing a voice command. Gamers Nexus editor-in-chief Stephen Burke called the findings “an egregious invasion of privacy.”
Beyond the microphone: network scanning
The investigation also alleges LG’s TVs actively scan local Wi-Fi networks, cataloguing other connected devices in a household — phones, smartwatches, routers, thermostats, air purifiers and PCs — along with network names, signal strengths and channel numbers. Researchers additionally flagged what they describe as a potential remote-code-execution vulnerability in the TV’s software. LG Ads Solutions, the company’s advertising arm, has previously said it can reach 363 million “addressable secondary devices” tied to LG televisions through Automatic Content Recognition, the technology that identifies what is playing on screen to serve targeted ads.
LG’s rebuttal
LG has pushed back firmly. In statements reported by outlets including The Register, TechRadar and Tom’s Hardware, the company said voice data is “processed only when the voice button on the remote control is pressed and held, or when a wake word such as ‘Hi LG’ is recognized,” and that outside of those instances its televisions do not collect or record ambient conversations. LG also said Automatic Content Recognition is opt-in and that claims of standby-mode eavesdropping "are not true." The dispute leaves consumers with dueling narratives: an independent technical teardown versus a manufacturer’s categorical denial, with no independent regulator yet weighing in on whose account is accurate.
Why this keeps happening to smart TVs
Privacy advocates note that smart TVs occupy a uniquely sensitive spot in the connected home — an always-powered, internet-linked device with a microphone and often a camera, sitting in the living room or bedroom. Unlike a phone, which users physically carry and consciously activate, a television is ambient furniture that many owners forget is listening at all. The LG dispute follows earlier allegations in July 2026 about LG’s monitor line collecting similar telemetry, suggesting the concerns are not limited to a single product category within the company’s lineup.
A skeptical read from security researchers
Some security professionals covering the story, including analysts at Malwarebytes, have cautioned that not every claim in the investigation has been independently reproduced, and that manufacturers’ voice-processing pipelines can create artifacts — such as brief post-wake-word buffering — that are easily mistaken for always-on recording without controlled testing. That caveat cuts both ways: it means the most alarming claims deserve scrutiny, but it does not resolve the core question of what data LG’s TVs collect by default and how long they retain it.
What it means for owners, and what’s next
For the tens of millions of households with an LG webOS television, the immediate practical step is checking the set’s voice-recognition and Automatic Content Recognition settings, both of which can typically be disabled in the privacy menu regardless of which account of the facts is correct. Longer term, the episode adds to a pattern seen this year across smart glasses and voice assistants: consumer AI features are increasingly bundled with data collection whose scope is disclosed, if at all, in dense settings menus rather than plain language. Whether LG faces a regulatory inquiry, a class-action suit, or simply a public-relations bruise will depend on whether independent labs can replicate the Gamers Nexus findings — and on how proactively agencies like the Federal Trade Commission choose to scrutinize AI-enabled home electronics going forward. Consumer advocates say the episode should push manufacturers toward clearer, plain-language disclosures at setup time, rather than relying on owners to discover privacy toggles buried several menus deep. Until then, the safest assumption for any AI-enabled television is that its microphone is a feature to actively manage, not one to forget about once the box is unpacked.
Photo: https://kaboompics.com/ / PEXELS via Pexels
