Reverse Engineering the Tizen OS to Uncover Samsung TV Telemetry
A cybersecurity researcher known online as log4j reported discovering extensive data-collection capabilities within Samsung’s Tizen operating system. The investigation began after noticing unusually targeted advertisements on a television purchased in 2025, according to trak.in. The researcher obtained root-level access by reverse-engineering the firmware, uncovering telemetry related to viewing habits, network activity, video playback progress, and advertisements skipped by users. These findings have not been independently verified on an unmodified retail television, and techradar.com reported that the situation requires caution before drawing conclusions.
Audio Categorization and the Family Care Feature
Among the code discovered during the firmware analysis were markers for dog-barking and baby-crying events, computer-vision processing, and a region-specific neural network designed to analyze viewing behavior. While these findings initially raised alarms about secret monitoring, techradar.com reported that the AI baby crying detection was actually part of Family Care, a feature Samsung actively advertised as a key selling point. The functionality was intended to send alerts to connected mobile devices when a child’s cries were detected. Samsung confirmed that the feature is being discontinued just 18 months after launch due to low usage and persistent false positives.
Evaluating Nearby Microphones and Advertising Data
The firmware analysis also identified code designed to detect external microphones in nearby connected devices, such as soundbars or Bluetooth hardware, even when the television itself lacked an integrated microphone. A second independent test conducted by a Hugging Face engineer known as ngxson used a Flipper Zero device to examine microphone behavior. That test observed the television opening a microphone for approximately five seconds before stopping, with no further activity detected over subsequent hours. However, trak.in noted that because the test television had already been rooted and several settings disabled, the results cannot establish the baseline behavior of an unmodified Samsung operating under normal conditions.
The Response from Industry Platforms and Security Researchers
The investigation drew further attention when log4j announced that OpenAI had banned their account for alleged cyber abuse. The researcher disputed the penalty, explaining that the tools were used solely for local reverse engineering rather than exploiting vulnerabilities, and suggested an automated classification system triggered the ban. Meanwhile, the broader implications center on how manufacturers combine automatic content recognition and advertising identifiers to personalize marketing. Users concerned about data collection can review their privacy, advertising, and viewing-information settings, or disable automatic content recognition features where available.
