A security researcher claims a 2025 Samsung smart TV uses extensive viewing data, device identifiers and local AI profiling to target adverts, though the findings have not been independently audited.
A security researcher who goes by the handle “log4j” says he found unusually precise targeted advertising on a 2025 Samsung smart TV and then dug into the set’s firmware. According to Tom’s Hardware, he used reverse engineering to gain root access and later posted his findings on X, where he described a system that appears to collect far more behavioural data than many viewers would expect. The alleged tracking is said to include watched videos, skipped ad breaks and playback progress, all feeding into ad targeting on Samsung’s Tizen-based platform.
The same account claims Samsung’s software also relies on a local artificial intelligence model to profile viewing habits, using a region-specific convolutional neural network to classify user interests. In addition, the system is said to make use of an advertising identifier for cross-device tracking, while also capturing screenshots and audio data for categorisation. The researcher further argues that even televisions without an integrated microphone may attempt to use connected peripherals, such as soundbars or Bluetooth accessories, to gather audio.
Samsung’s own privacy policy confirms that its visual display products may process a wide range of device information, including connected-device details, software versions, unique identifiers and advertising IDs for Samsung smart TVs. The company’s support materials also say consumers can change privacy settings, including controls linked to Automatic Content Recognition and smart TV identifiers. That does not resolve the broader concern, but it does show that the company openly describes at least some of the tracking architecture used on its televisions.
The dispute also lands in a market where smart TV privacy has been under pressure for years. Research from RTINGS has found that modern televisions often function as advertising platforms as much as display devices, with network traffic that is difficult to separate from telemetry and recommendation systems. Ars Technica recently reported on fresh research into automatic content recognition across multiple smart TV brands, underscoring how widespread the issue has become. LG has faced similar criticism in recent months, with the company defending its television data practices and saying that features such as voice recognition, interest-based advertising and ACR are separate options that can be disabled.
For now, the claim remains one researcher’s technical analysis rather than an independently audited disclosure from Samsung. Even so, it adds to an increasingly familiar pattern: connected TVs are cheap in part because they are also data-gathering devices. If the findings are confirmed in full, the implications would reach beyond advertising. They would raise further questions about the balance between smart features, background processing and the performance penalties users often complain about on modern TV interfaces.
Disclaimer: This content is intended for informational purposes only. Readers are advised to exercise their own judgement, conduct due diligence, or consult a qualified expert before acting on any information provided.
