LG Smart TV Privacy Claims Trigger a 216 Million Device Dispute
LG has rejected explosive LG smart TV privacy claims involving 216 million televisions, alleged audio recording, network scanning, and extensive data collection.
The allegations come from a 135-minute Gamers Nexus investigation published on September 6, 2026. Researchers examined network traffic and firmware behavior on several U.S. retail LG televisions. They claim the TVs logged activity, identified nearby devices, sampled displayed content, and retained voice-related data under conditions owners would not expect.
LG says the central spying allegations are false. It told Tom's Hardware that its televisions do not record ambient conversations unless users intentionally activate voice functions. However, the company confirmed that enabled televisions can monitor locally for the “Hi LG” wake word while in standby.
That disagreement matters because it reaches beyond one microphone or one television model. It tests whether consumers can understand and control the data generated by a connected screen inside their homes.
The investigation also raises a harder question about verification. Gamers Nexus presented packet captures and firmware analysis, but its broad conclusions have not received independent replication. LG provided categorical denials, yet its response did not address every technical allegation.
The story is therefore not a settled discovery of 216 million spying televisions. It is a high-stakes dispute between an independent technical investigation and a manufacturer whose advertising business depends partly on television data.
LG Smart TV Privacy Claims Meet a Direct Denial
The immediate change is that LG has now answered the investigation, narrowing the dispute to specific technical and consent questions.
Gamers Nexus worked with security researchers known as MrBruh and uturn. Their security investigation used packet capture and firmware analysis to examine how selected LG televisions behaved.
Packet capture records data moving across a network, including the destinations contacted by a device. Firmware analysis examines the software embedded inside the television and the functions that software can perform.
The investigators claim they observed persistent logging and data collection across several operational states. Those states reportedly included normal use, offline periods, and standby, when the screen appeared to be off.
They also reported evidence related to audio processing. Their most serious allegation is that LG televisions captured or stored microphone information beyond the limited conditions described to users.
Gamers Nexus further claimed that some voice transcripts were stored in plain text. Plain text means readable information that lacks an additional encryption layer within the examined storage context. Tom's Hardware noted that LG did not address this particular allegation in its response.
The report also described televisions identifying devices on the local network. Those devices reportedly included phones and printers, creating concern that a television could help map activity across an entire household.
LG disputes the interpretation of those findings. In its LG response, the company said the video’s claims “are not true.”
LG says voice data is processed only after a user holds the remote’s voice button. Processing can also begin when an enabled television recognizes the “Hi LG” wake phrase.
The company says other ambient conversations are neither collected nor recorded. It also says unrecognized audio remains on the device, receives local processing, and is immediately deleted.
That explanation acknowledges an important distinction. An enabled television must analyze incoming sound locally to determine whether someone said the wake phrase. Local wake-word detection is not automatically equivalent to uploading household conversations.
However, local processing still means that a microphone and software remain active enough to evaluate sound. The privacy question becomes what the television retains, when it transmits data, and whether its controls match LG’s description.
LG also confirmed local network scanning. It says televisions identify compatible devices to support connections, content sharing, and smart-home functions.
Finding another device on the same network is not inherently malicious. Streaming boxes, speakers, televisions, and home automation hubs routinely use discovery protocols to locate compatible hardware.
The concern arises when discovery supports profiling that users did not reasonably anticipate. A connection feature and an advertising identifier can examine similar network signals while serving very different purposes.
Neither Gamers Nexus nor LG has independently settled that distinction. Tom's Hardware expressly said it had not verified the investigation or LG’s counterclaims.
That verification gap should define coverage of the story. Calling every LG television a proven spying device exceeds the available evidence. Treating network scanning and standby audio processing as meaningless would also dismiss legitimate questions.
The 216 Million Figure Is a Claim, Not a Test Result
The headline number describes the alleged reach of LG’s television platform, not 216 million individually tested devices.
Gamers Nexus titled its investigation “216,000,000 Spy TVs.” The scale makes the allegations memorable, but it can also obscure what the researchers actually examined.
The investigation involved selected retail televisions rather than a forensic audit of every deployed LG model. Findings from those units can reveal behavior worth investigating, but they do not establish identical behavior across all regions and software versions.
Televisions sold in different years can use different processors, microphones, firmware, privacy menus, and advertising components. Regional regulations can also change which systems operate and where data travels.
The 216 million figure should therefore be read as a platform-scale framing. It is not evidence that researchers observed 216 million microphones recording conversations.
This distinction does not make the research unimportant. A reproducible behavior on a smaller sample can identify a systemic design issue, particularly when shared software reaches many products.
It does change the next evidentiary step. Independent researchers need to reproduce each major observation across documented models, firmware versions, account states, consent settings, and network conditions.
Researchers must also separate ordinary telemetry from content recognition and voice processing. Telemetry usually reports device operation, errors, or feature use. Content recognition identifies displayed media, while voice systems process microphone input.
Those systems can produce network requests that look suspicious without payload inspection. Conversely, an innocent-looking encrypted connection can carry sensitive information that a basic destination list cannot reveal.
The investigation’s packet captures offer one layer of evidence. Firmware findings can add context by showing which functions generated records or called particular services.
Still, a packet’s existence does not alone reveal its meaning. Researchers need repeatable tests that connect a user action, local record, transmission, destination, and documented purpose.
LG’s categorical denial does not remove this need. It increases it. The company now has a specific public account that outside researchers can test.
For voice privacy, the crucial experiment is straightforward in concept. Researchers can enable and disable far-field recognition, place the television in standby, present controlled speech, and monitor local storage and network traffic.
A strong test would include phrases that resemble the wake word and conversations that do not. It would document whether any audio, transcript, or derived identifier remains after each condition.
The same discipline applies to local network scanning. Researchers should identify the protocols used, devices discovered, data recorded, and servers receiving any related information.
A television can discover a phone for screen casting without sending the phone’s identity to an advertising service. If it links that phone to an audience profile, the privacy implications become substantially different.
LG says its network discovery serves connectivity, sharing, and smart-home functions. That is plausible, but the company has not publicly answered every reported observation about how those discoveries are stored or used.
The investigation also connected household device visibility to statements from LG’s advertising operation. Marketing language about knowing which devices exist in an LG household can sound broader than a simple casting feature.
Such language does not prove that a specific television secretly uploaded a complete network inventory. It does show why consumers and researchers want a precise technical explanation.
LG can reduce the uncertainty by publishing data-flow documentation. That material should identify each process, collected field, retention period, destination, and controlling consent switch.
Without that detail, both sides retain room to interpret the same behavior differently. Gamers Nexus sees a data dragnet, while LG describes ordinary connected features governed by user choices.
Consent Is the Real LG Smart TV Privacy Test
The core conflict is not whether a smart television processes data, but whether users knowingly authorize each purpose and can stop it.
LG says Automatic Content Recognition, voice recognition, and interest-based advertising require separate, explicit consent. It also says users can disable those functions through television settings.
Automatic Content Recognition, or ACR, creates fingerprints from displayed video or audio and matches them against a reference library. This can identify programming from broadcasts, streaming services, game consoles, or other connected sources.
ACR supports recommendations and audience measurement. It can also give advertisers a detailed account of what appears on a television, even when the content comes through HDMI.
LG’s current privacy policy directs smart television owners to a separate Smart Media Product Membership Privacy Policy. The applicable terms can vary with location, model, and software.
That fragmented structure creates a usability problem. Consumers rarely distinguish account data, viewing information, voice information, device discovery, and interest-based advertising while configuring a television.
LG says each sensitive feature uses an optional agreement. Its support guidance also says agreements covering viewing information, voice features, and personalized advertising are unnecessary for basic smart services.
The unresolved issue is whether those choices remain clear during actual setup and later software updates. A formally optional agreement can still create confusion if declining it disables an expected feature or produces repeated prompts.
Consent also needs to control the underlying technical behavior. A disabled setting should stop the related collection and transmission, not merely prevent personalized advertising after the data reaches a server.
There is useful independent evidence on this narrower question. A peer-reviewed ACR study examined LG and Samsung televisions in the United States and United Kingdom.
The researchers found that ACR could operate while televisions served as external displays through HDMI. That result matters because some owners use smart televisions without opening built-in streaming applications.
The study also found that opting out stopped communication with the identified ACR domains on the tested televisions. The researchers concluded that the examined opt-out mechanisms worked for ACR network traffic.
That finding does not resolve the 2026 audio allegations. It does provide an important counterweight to any blanket claim that privacy controls never affect LG television tracking.
It also shows why the current controversy should be divided into testable components. ACR traffic, wake-word processing, device discovery, diagnostic logs, and advertising profiles require separate evidence.
One working control does not prove that every control works. Likewise, one concerning log does not prove that every television continuously uploads household conversations.
The geographic findings add another complication. The ACR researchers observed regional differences between U.S. and U.K. televisions, including different domains and behavior for some viewing scenarios.
Results from a U.S. retail model cannot automatically describe every LG television worldwide. Any claim tied to 216 million devices must account for those regional variations.
The same applies to firmware versions. An update can add, remove, or change a data flow without altering the television’s physical hardware.
Users also need to know what disabling a feature actually means. Turning off personalized advertising might stop targeting while leaving audience measurement active. Disabling voice recognition might differ from disabling far-field wake-word detection.
LG’s statement presents these as separately controlled services. Independent testing should verify the boundaries between them, including whether historical data remains after a feature is disabled.
The burden does not rest entirely on consumers. A television interface should translate complex data practices into choices that match ordinary expectations.
“Voice recognition” should state whether audio leaves the device, when transmission starts, and how long any transcript remains. “Viewing information” should explain whether HDMI content receives fingerprints.
“Nearby devices” should disclose whether discovery results remain local. If identifiers enter an advertising profile, that separate purpose should require a clear and specific choice.
Consent becomes meaningful only when the labels, behavior, and consequences align. That alignment is now the most important test for LG’s defense.
Smart TV Tracking Has Already Triggered Enforcement
The industry’s history shows that undisclosed viewing surveillance can move quickly from a product feature to a regulatory case.
LG is not the only television company to use viewing data or ACR. Samsung and Vizio have also operated connected television platforms supported by advertising and audience measurement.
The relevant comparison is not which manufacturer collects any data. Nearly every modern connected platform generates operational and account information.
The meaningful comparison concerns informed consent, collection scope, data sharing, and whether users can disable tracking without losing core functionality.
The United States already has a major precedent. In 2017, the Federal Trade Commission and New Jersey announced a settlement with Vizio over viewing-history collection.
The agencies alleged that Vizio captured second-by-second information from 11 million televisions. According to the complaint, the data covered cable, broadband, set-top boxes, DVDs, broadcasts, and streaming devices.
Regulators also alleged that demographic information was appended to the viewing histories. Third parties reportedly used the combined information for advertising and audience analysis.
The dispute focused heavily on disclosure. Vizio described its “Smart Interactivity” feature as enabling offers and suggestions, according to the FTC.
Regulators said that description did not adequately tell consumers about viewing-data collection. The resulting Vizio consent order required prominent disclosure and affirmative consent.
The settlement also required deletion of certain previously collected data and a comprehensive privacy program. Vizio agreed to a total payment of $2.2 million, including a suspended portion.
That case does not establish that LG violated any law. It does show the standard regulators have previously emphasized for connected television tracking.
A technically available opt-out may not be enough if the surrounding description misleads users. A recommendation label cannot quietly cover detailed viewing surveillance.
The LG dispute presents an even more sensitive allegation because it includes microphones. Viewing histories reveal interests and routines, while ambient conversations can contain private speech from everyone near the screen.
LG firmly denies collecting such conversations outside intentionally activated voice features. That denial creates a factual question that controlled testing can address.
Network scanning occupies a less dramatic but still important category. Connected devices often need discovery services, yet a household device map can become valuable advertising information.
The risk depends on purpose and data movement. Local discovery for casting is different from persistent household profiling, even when both begin with the same network observation.
The advertising context makes clarity essential. Connected television businesses can generate revenue after the hardware sale through promotions, sponsored content, audience measurement, and targeted campaigns.
Those incentives do not prove misconduct. They do create pressure to collect more behavioral signals and connect television activity with other household devices.
Consumers have limited visibility into that process. A phone browser can display trackers through privacy tools, but a television often hides its connections behind an appliance interface.
Encrypted traffic makes inspection harder. Owners may see that a server received data without knowing whether the payload contained an error code, content fingerprint, device identifier, or voice record.
Manufacturers consequently hold more information than customers about the system’s operation. That imbalance makes precise documentation and credible independent audits especially important.
Competitors face the same pressure. If one platform reduces collection while another builds richer advertising profiles, the second platform might offer marketers greater reach.
Privacy rules can prevent that competition from becoming a race toward less visible tracking. Clear consent standards also give manufacturers a common baseline.
The 2024 ACR research found tracking on both LG and Samsung platforms, though implementation and traffic patterns differed. The authors reported that Samsung sent up to twice as much data to ACR domains in their tested configurations.
That result should not become a simplistic privacy ranking. Data volume alone does not show sensitivity, necessity, retention, or downstream sharing.
It does confirm that ACR is an industry system rather than an isolated LG invention. The present controversy pressures every smart television maker to explain how its screen becomes an advertising sensor.
What LG’s Denial Does Not Yet Resolve
LG answered the ambient-recording accusation, but several technical and evidentiary questions remain open.
The company’s clearest statement concerns wake-word processing. LG says users must manually activate far-field voice recognition before the television listens for “Hi LG.”
When enabled, the system monitors for that phrase while the television is in standby. LG says the sound receives local processing and immediate deletion if the wake word is absent.
This account is technically coherent. Many voice assistants use on-device processing to identify a wake phrase before initiating a server request.
However, the claim still requires verification on the models and firmware examined by Gamers Nexus. Researchers need to confirm what “immediately deleted” means in logs, temporary storage, crash records, and diagnostic buffers.
The company did not fully address the reported plain-text transcripts. It also did not publicly provide a line-by-line response to each vulnerability described in the investigation.
A transcript’s presence would not automatically prove continuous remote surveillance. It could result from an intentionally activated voice command, debugging function, or local feature.
Its storage format and retention still matter. Readable transcripts could expose sensitive commands if another account, application, or vulnerability gained access to the television.
The security allegations require a separate analysis from LG’s intended data practices. A feature can follow its privacy policy yet remain exploitable by an attacker.
Gamers Nexus claims the tested televisions contained vulnerabilities that could turn built-in functions into covert listening tools. That is not the same allegation as saying LG intentionally operates those tools for surveillance.
Coverage should preserve that distinction. Corporate data collection, insecure local storage, and malicious exploitation describe three different threat models.
LG says it continuously monitors its platform, works with independent security experts, and supplies updates. It did not provide technical details about the reported vulnerabilities in the published response.
Responsible disclosure timelines can limit what a company says before patches reach customers. Silence can therefore reflect remediation work, but it cannot serve as proof that remediation is underway.
The network-scanning admission also needs more detail. LG describes it as a standard capability for finding compatible devices and enabling content sharing.
The company should identify which discovery methods operate by default, which identifiers they capture, and whether results leave the local network. It should also disclose retention and advertising uses.
The investigation’s broader logging claims remain similarly unsettled. Connected televisions must create some logs to diagnose errors and maintain services.
The relevant questions concern content, duration, access, transmission, and user control. A timestamped error message carries different risk from a searchable record of household activity.
The 216 million framing presents another unresolved issue. Public reporting has not established that every unit included in that figure contains the same components or behavior.
A precise response from LG could list affected and unaffected model families. It could also specify regional differences, minimum firmware versions, and the features enabled by default.
Gamers Nexus bears responsibilities as well. Its most consequential claims need reproducible procedures, clear test conditions, and enough technical material for independent examination.
A long video can reveal extensive evidence, but length does not substitute for replication. Security researchers should be able to recreate each claimed behavior without relying on interpretation alone.
The best outcome would be a model-by-model validation matrix. It should compare default settings, optional agreements, standby behavior, local files, network destinations, and user-initiated voice actions.
Independent laboratories could then repeat the tests. Results might support LG, confirm parts of the investigation, or show that both accounts describe different configurations.
Until that happens, consumers face an uncomfortable middle ground. The central accusations are serious, but the available evidence does not justify treating every claim as conclusively established.
LG’s denial also cannot close the issue by itself. The company operates the platform under examination and has the greatest access to its internal data flows.
The right standard is not trust in a manufacturer or a media channel. It is reproducible evidence supported by precise documentation and independently observable behavior.
Three Signals Will Determine What Happens Next
The controversy will turn on independent replication, LG’s technical response, and any regulatory demand for evidence.
The first signal is whether outside researchers reproduce the audio findings. This should happen across multiple LG models, regions, firmware versions, and consent configurations.
Replication that finds unintended audio retention or transmission would strengthen the investigation’s central warning. It would also force LG to explain whether the behavior reflects a bug, design choice, or compromised feature.
Failure to reproduce the result under documented conditions would weaken the broad spying claim. It would not eliminate every security concern, but it would narrow the dispute significantly.
Researchers should publish exact state transitions. Tests need to distinguish a powered screen from standby, a disconnected router from stored credentials, and disabled voice control from enabled far-field listening.
The second signal is a detailed LG security advisory or firmware update. A versioned advisory would show that LG has investigated specific code paths rather than issuing only a general denial.
An update addressing transcript storage, microphone permissions, logging, or network discovery would validate at least part of the technical concern. It would not necessarily validate claims about intentional surveillance.
If LG concludes that no fix is required, it should publish supporting technical evidence. That evidence should map consent settings to actual processes and network destinations.
The company could also commission a credible independent audit. A useful audit would identify its scope, tested products, limitations, and findings instead of offering a general certification.
The third signal is regulatory interest. Consumer protection authorities can demand records that external researchers cannot access, including internal design documents and advertising data flows.
Regulators would likely focus on whether disclosures match behavior. The Vizio precedent shows that affirmative consent and accurate descriptions matter when televisions collect viewing histories.
An investigation would not establish wrongdoing. It would indicate that the verification gap has become significant enough to justify formal information requests.
Absence of immediate enforcement would not clear LG either. Agencies often need time to assess complaints, technical findings, and jurisdictional questions.
Owners do not need to wait passively. They can review the television’s user agreements and disable optional viewing information, voice recognition, and personalized advertising settings.
Users who do not need built-in smart functions can also keep the television offline and use a separate streaming device. That choice changes the data controller, rather than eliminating connected-device tracking.
A separate device can still collect viewing and account information. It may offer clearer controls, faster updates, or an easier replacement path, depending on the product.
Owners should also install verified security updates. Avoiding all updates can leave known vulnerabilities unpatched, even when concerns about new terms are understandable.
The larger lesson concerns observability. A connected television now combines a screen, microphone, operating system, advertising platform, and local-network presence.
Consumers cannot evaluate that combination through picture quality reviews alone. Privacy controls and support lifetimes belong in the buying decision alongside brightness, ports, and software features.
LG can strengthen its position by answering the unanswered technical claims with evidence. Gamers Nexus can strengthen its case through reproducible artifacts and independent confirmation.
For now, the LG smart TV privacy claims remain a documented controversy rather than a final verdict. The investigation identified questions that deserve testing, while LG supplied denials that can themselves be tested.
Watch for replicated microphone results, a model-specific LG advisory, and regulatory requests for internal records. Those three developments will show whether this becomes a security incident, a disclosure dispute, or an overstated interpretation.
