Modern televisions are advertising platforms as much as displays. Most can identify what appears on screen and report it back, a practice known as automatic content recognition, and it is an industry-wide norm rather than one brand’s quirk.
The technology behind the screen
A smart television is a display with a computer attached: an operating system, an app store, a permanent internet connection and, in most cases, a microphone somewhere in the remote control. That computer does several things at once. It streams video. It runs apps. And, on most sets sold in the past decade, it also runs software that tries to work out what is being shown on the panel.
The technical name for that last function is automatic content recognition, usually shortened to ACR. It works by sampling the picture, and sometimes the audio, at intervals, reducing each sample to a compact numerical fingerprint, and comparing that fingerprint against a reference library of known content. If the fingerprint matches, the television knows it is showing a particular programme, film, advertisement or channel. Crucially, this happens regardless of where the picture came from: a streaming app, an aerial, a games console or a set-top box plugged into an HDMI port. The television does not need to understand the source. It only needs to see the pixels.
From display to advertising platform
The commercial logic behind ACR is not mysterious. Flat panels became a commodity, manufacturing scale moved to a handful of suppliers, and the profit margin on the hardware itself compressed towards nothing. Selling a television once, at a thin margin, is a far worse business than selling a television and then continuing to earn from it for years.
Connected platforms provided the answer. Every major manufacturer now ships its own operating system with its own home screen, its own app storefront and, increasingly, its own free ad-supported streaming channels. The home screen is prime advertising inventory. The storefront takes a cut of subscriptions. The free channels carry advertising sold by the manufacturer. All of this is worth more if the platform knows what the household actually watches, which is precisely what ACR supplies.
The result is that television makers became advertising companies that also sell hardware. This shift happened gradually and with little public discussion, which is part of why periodic revelations about it land as a shock.
The data pipeline today
In practice, the viewing data collected by a television is rarely useful on its own. Its value comes from being joined to other things.
The set typically reports fingerprints or match results to a server operated by the manufacturer or a specialist ACR partner, along with identifiers such as the device’s advertising ID and its internet address. That internet address is shared with every other device in the home, which is what allows viewing behaviour on the television to be linked to browsing and app activity on phones and laptops in the same household. Advertising firms maintain identity graphs for exactly this purpose. The pitch to advertisers is that a person who saw a commercial on the living-room screen can then be shown a related advertisement on their handset.
Consent for this usually exists, formally. It is generally collected during initial setup, in a screen that appears alongside the terms of service and network configuration, under a label such as viewing information services or similar wording that does not obviously describe content tracking. The option can normally be declined or switched off afterwards in the settings menu, though the location and naming vary by manufacturer and by firmware version.
The Verge reports that this broader industry picture came back into view after a lengthy video from the outlet Gamers Nexus alleged that LG televisions record and store audio even when they appear to be switched off, track everything watched, and could in principle be compromised remotely and used for covert surveillance. The Verge’s framing is that the underlying behaviours are not confined to a single manufacturer. The precise technical accuracy of each individual claim in that video is contested and is not something this article can independently confirm.
Points of confusion
The most common error is treating this as a single company’s scandal. Content recognition, home-screen advertising and cross-device identity matching are standard features of the connected television business across the major platforms. Singling out one brand tends to imply that switching to a competitor solves the problem, which it generally does not.
A second confusion is between microphones and content tracking. These are separate systems. ACR reads the screen and does not require a microphone at all. Voice assistants use a microphone, usually in the remote, and normally transmit audio only after a wake word or a button press. Claims that a set records continuously while apparently off are claims about the microphone system specifically, and they require separate evidence from claims about viewing data.
Third, many people assume that using a television only as a monitor, with an external streaming stick or console, avoids the issue. Because ACR operates on the displayed image, it can in principle recognise content arriving over HDMI as well.
Finally, there is the belief that simply never connecting the set to the internet is a complete answer. It is effective, in that data cannot be transmitted, but it also disables apps, security updates and in some cases parts of the interface. Whether any particular model behaves differently when offline over long periods is not publicly documented in any consistent way.
Where to look next
The most direct source is the television in the room: the privacy or general settings menu, where content recognition controls, advertising identifiers and voice features are usually listed separately and must each be turned off individually. Manufacturers’ published privacy policies describe the categories of data collected, though typically in broad language.
Beyond that, consumer organisations in several countries have published comparative assessments of smart television privacy settings, and academic network measurement groups have run studies that capture the traffic leaving these devices. Competition and consumer protection regulators in the United States and the European Union have both examined connected television advertising, and their published materials are useful for understanding the commercial structure rather than the engineering. For the current dispute specifically, the original video and the manufacturer’s response are the primary documents.
Frequently asked questions
What is automatic content recognition on a TV?
Automatic content recognition is software built into most smart televisions that identifies what is being displayed. It samples the picture or audio, converts each sample into a numerical fingerprint, and matches that against a reference database of known programmes and advertisements. The result tells the platform what the household is watching, which is then used to target advertising and to measure audiences.
Can a smart TV record audio when it is switched off?
That specific allegation has been raised about certain sets, and The Verge reports it as one of the claims made in a Gamers Nexus video about LG televisions. Whether it holds for any given model is contested and not independently established here. What is documented generally is that microphones in modern televisions usually sit in the remote control and transmit after a wake word or button press.
Does disconnecting a TV from the internet stop the tracking?
Keeping the set permanently offline prevents it from transmitting viewing data, because the data has nowhere to go. The trade-off is that apps, firmware updates and some interface features stop working, so the television becomes a plain display driven by an external device. That external device, however, will have its own data collection practices.
Do all television manufacturers collect viewing data?
Content recognition and advertising identifiers are standard across the major connected television platforms, which is the central point in The Verge’s coverage. Implementation details, default settings and the wording used to request consent differ between manufacturers and between firmware versions. No public comparison establishes a definitive ranking of which platform collects the least, so switching brands is not a reliable fix.
How is smart TV viewing data used by advertisers?
Viewing data is typically combined with the household’s internet address and advertising identifiers, then matched against identity graphs maintained by advertising firms. This allows a household that saw a particular commercial on television to be shown related advertising on phones, tablets and computers connected to the same network. It also supports audience measurement and the pricing of advertising slots on the platform itself.
Sources and further reading
- The Verge — technology news coverage of connected television data collection and the current dispute over LG sets.
- Gamers Nexus — the independent hardware channel whose long-form video prompted the renewed discussion.
- Consumer protection and competition regulators in the United States and European Union — published examinations of connected television advertising and consent practices.
- Academic network measurement research groups — studies that inspect the traffic leaving smart televisions under laboratory conditions.
Surfaced from the rss:verge signal “smart TV privacy claims”. AI-assisted draft, editorially reviewed.
