The remote has one button everybody understands. Press it, the picture disappears, and the television stops being the centre of the room. The conversation continues. Someone puts the kettle on. The cat resumes whatever management position it held before the film started.
“Off” is a wonderfully uncomplicated word for an increasingly complicated appliance.
This week, LG’s televisions became the subject of a much less comfortable conversation. A Gamers Nexus investigation raised concerns about viewing-data collection, network discovery and audio handling. LG disputed the central allegations. By September 12, privacy specialists were weighing in on what the company’s response did—and did not—explain.
It is tempting to treat this as a contest between a frightening headline and a reassuring statement. Pick the side you trust, change a setting, move on.
But there is a more useful question hiding underneath: why is understanding the behaviour of a television now a research assignment?
Smart TV privacy should not depend on a viewer knowing which part of a device is listening, which part is recognizing a programme, which part is finding a nearby speaker, and which part is selling an advertiser a better audience. Those are different functions with different consequences. A person buying a screen should be able to make separate, understandable decisions about them.
The real failure is not simply that connected televisions collect information. It is that the boundary between using the appliance and participating in its business model can be so difficult to see.
Four different questions, one frightening headline
The Gamers Nexus video, published September 6, combined an investigation of LG’s first-party behaviour with security research. Those categories should remain separate. A capability demonstrated after a device is compromised is not, by itself, proof that an ordinary, uncompromised television routinely performs the same action for its manufacturer.
LG’s response, published by Tom’s Hardware on September 9, says voice processing occurs when someone holds the remote’s voice button or triggers an enabled wake-word feature. It says wake-word detection happens locally, with unrecognized audio deleted rather than transmitted. LG also acknowledges local-device discovery, describing it as a smart-home function, and says automatic content recognition is opt-in. Tom’s Hardware explicitly says it has not independently verified either side’s claims.
DigitalNights has not independently tested these televisions either. Claims that LG routinely records household conversations should not be treated here as established fact.
That leaves four questions worth asking without turning them into one enormous accusation.
What does the television recognize on its screen? When does its microphone operate? Which other devices does it discover? What could happen if its software were compromised?
An answer to one does not settle the others. A legitimate reason to discover a wireless speaker does not explain the retention of discovery data. A locally processed wake word is not the same as sending a conversation to a server. Recognizing a film is not equivalent to understanding everything someone types on a connected laptop.
Precision matters because an exaggerated claim gives a company an easy target. It can refute the most dramatic interpretation while a less cinematic, thoroughly legitimate concern remains unresolved.
There is also a fairness issue. Security researchers deserve to have the conditions of their demonstrations understood. Manufacturers deserve to have their responses reported accurately. Readers deserve something better than being asked to choose a team before anyone explains the experiment.
The screen has a second audience
Automatic content recognition, usually shortened to ACR, sounds like a feature that might help you remember the name of an actor. Its commercial purpose can be considerably broader.
LG Ad Solutions’ own technology page describes viewing information used to build advertising audiences. Its examples include programmes, movies, advertisements, games and streaming apps, alongside targeting categories such as viewing habits, service use and location. This is the company describing its product to prospective customers, not a critic describing a hypothetical scheme.
In that relationship, the viewer and the advertiser are interested in the same screen for very different reasons. One wants to know whether there is anything decent on. The other wants to know what kind of household is watching.
That does not mean all advertising is objectionable, or that every use of viewing information is automatically abusive. It means the television has a second commercial role which deserves to be as visible as its first.
A person may reasonably understand that a streaming service remembers shows watched inside its app. The more surprising possibility is that the display itself can participate in recognizing content delivered by another device.
A UCL research summary published November 10, 2024 describes tests of Samsung and LG televisions in the UK and US. Researchers observed ACR-related activity when televisions displayed content from external devices, not only conventional television. They also reported that opting out stopped the observed ACR activity. Importantly, encrypted traffic limited their ability to inspect the exact information transmitted.
That study is not a verdict on every model, firmware version or country in 2026. It is evidence that “I only use it as a screen” and “the screen cannot contribute viewing information” are not interchangeable assumptions.
It is also a reason to take controls seriously rather than telling everyone they are useless. An opt-out that changes observed behaviour matters. The next question is whether an ordinary owner can find it, understand its scope and tell that it worked.
The historical context makes that question harder to dismiss. In February 2017, the US Federal Trade Commission announced a $2.2 million Vizio settlement over allegations of collecting viewing histories without consumers’ knowledge or consent. The order required prominent disclosure and affirmative express consent, among other measures. That was a different company and a different case; it is not evidence that LG violated the same rules.
It does show how long the industry has had to get better at explaining this arrangement.
Nine years later, a household should not still need to translate advertising-industry language to work out the basic bargain. Picture quality is explained in colourful diagrams. Viewing-data collection ought to be explainable without a glossary.
Consent needs an understandable result
The usual defence is that people agreed. The usual reply is that nobody reads the agreements. Neither gets us very far.
The useful test is what the person could reasonably understand at the moment of choosing. A button labelled with a pleasant phrase about improving recommendations is not necessarily an adequate explanation of household profiling. Equally, a setting labelled “limit advertising” should not be assumed to stop every kind of collection or every network connection.
In TechRadar’s September 12 interviews, UCL researcher Anna Maria Mandalari highlighted the difference between saying information is not used for advertising and explaining its entire lifecycle. The article also reported that her testing found ACR reporting to advertisers stopped after opting out. Both points matter: acknowledge a working control, and ask precisely what it controls.
Here is a practical standard DigitalNights would apply to any connected television. This is an editorial test, not a certification programme or a claim about the present menus on a particular LG model.
For every optional data feature, the manufacturer should give an owner four plain answers: what is collected, where it goes, how long it remains, and what changes when permission is withdrawn.
The last answer is often the most important. It converts an abstract promise into a consequence someone can understand.
“Your viewing history will no longer be sent for advertising measurement” is different from “ads will be less relevant.” “Voice commands will be unavailable” is different from “the microphone is electrically disconnected.” “Previously collected information will be deleted” is different from “future collection will stop.” These are illustrative statements, not quotations from a manufacturer’s settings.
A useful privacy screen would make those distinctions explicit. It would also distinguish the television’s own behaviour from the behaviour of installed apps. One switch cannot honestly promise to govern a different company’s account and servers unless the product actually has that authority.
This is where the problem becomes a design issue rather than an argument about how diligent consumers should be. Imagine changing a picture setting and being told only that your visual experience may become less optimal. You would quite reasonably ask whether the screen had become darker, blurrier or upside down.
Privacy settings deserve the same respect for cause and effect.
There is a credible counterargument: modern televisions do useful things that require connections and information. Voice control can be an accessibility feature. Device discovery can make casting less frustrating. Recommendations can reduce the nightly ritual of scrolling through sixteen rows before watching something familiar.
An offline-only television is not automatically the best television for everybody. Nor does useful software become sinister merely because it talks to a server.
The answer is not to pretend those benefits are worthless. It is to separate them from optional commercial uses wherever possible. Someone should be able to want a working remote microphone without being presumed enthusiastic about every other form of household measurement.
Manufacturers might object that explaining all these distinctions would overwhelm people. That is a reasonable concern about presentation, not a reason to leave the distinctions unexplained. A short summary can lead to deeper detail. The owner should not have to choose between a cheerful slogan and a legal document long enough to qualify as evening reading.
Privacy should survive ordinary use
For a household deciding what to do now, the most useful starting point is a modest one: identify the television’s exact model and software version, then consult its current privacy and voice-control documentation. Look for separate explanations of viewing recognition, advertising personalization and microphone behaviour. Do not assume one advertising toggle governs all three.
Write down what each setting claims to change before changing it. That small habit helps distinguish an actual assurance from an interpretation. If the documentation does not explain whether a control stops collection, sharing or merely personalization, that unanswered question is itself useful information—not proof of misconduct, but a limit on what the control promises.
This is not a model-specific walkthrough. Menu names and available features vary, and these steps do not certify a television as private. They are a way to make a more informed choice without pretending the living room has become a testing laboratory.
Using an external player while keeping the television off the network can separate the display from its online services. But it also removes built-in online functions, requires a plan for legitimate firmware updates, and shifts some privacy decisions to the player and its apps. Buying another box does not buy an exemption from data collection.
Historical research underlines that last point. The 2020 paper “The TV is Smart and Full of Trackers” examined advertising and tracking on Roku and Amazon Fire TV, as well as household television traffic. Its blocklist testing found both missed tracking services and broken functionality. It is an older study, not a current product ranking, but it offers a useful warning against treating a network filter as a universal privacy switch.
For enthusiasts, technical controls can still be worthwhile. The problem is making enthusiast-level maintenance the price of ordinary boundaries. A solution that requires someone to become the family network administrator may be satisfying to that person and completely impractical for everyone else.
The household deserves options that remain understandable when the technically confident person is away. It deserves to know what a software update changed. And it deserves an explanation of standby that does not rely on the comforting assumption that a black screen means every subsystem is inactive.
What should happen next is less glamorous than another viral exposé. Independent testing should specify the model, firmware, region, consent choices and whether the device was modified. Manufacturer responses should address those same conditions. Claims about routine collection should be separated from exploit demonstrations, and fixes should identify which problem they fix.
For product reviews, this suggests an additional question beside brightness, motion handling and gaming latency: how clearly can an owner understand and control the television’s secondary uses? Not a vague privacy score, but documented choices with documented consequences.
There is room for a manufacturer to compete on that clarity. A genuinely useful “display only” mode, an intelligible account of data retention, or a clear notice when an update changes optional features would be concrete product improvements. They would also be easier to evaluate than another promise that customer trust is important.
None of this requires deciding that every allegation against LG is true. The strongest argument survives uncertainty about the most dramatic claims: people should be able to understand the appliance they brought home.
The remote does not need another button. The one marked “off”—and all the less obvious switches behind it—needs an explanation worthy of the person holding it.
