@charlesgillespi - Samsung TV Audio Telemetry

Open original

@charlesgillespi on X A post by Charles Gillespi, quoting a report from @log4j, claims Samsung devices classify ambient audio on the device before sending results back to Samsung. The telemetry reportedly tracks sensor preferences, dog-barking and baby-crying event counts, and how long cameras stream, and the author calls it a serious privacy violation that Samsung should face criminal charges over.

  • Samsung devices reportedly classify what audio they hear locally before reporting results to the company.
  • Telemetry paths reportedly include counters for dog barking and baby crying events.
  • Camera streaming duration and sensor preferences are also reportedly logged and sent back.

The post is a short reaction from Charles Gillespi to something a researcher named log4j reported: that a Samsung device classifies the audio it picks up and then sends results back to Samsung. According to the quoted claim, the telemetry includes sensor preferences, counters for dog-barking and baby-crying events, and how long the camera has been streaming. Gillespi calls it one of the most flagrant privacy violations he has seen and asks for Samsung to be "eviscerated" and for criminal charges. The post itself contains no device model, firmware version, test setup, or independent verification, so everything beyond the quoted claim is unconfirmed. That gap matters. The strong language is Gillespi's opinion, not a finding. What the quoted material actually establishes is narrower and still worth checking: whether the device sends derived audio labels, such as "dog barking," as opposed to raw audio, and whether users were told and could opt out. Those two questions separate a disclosure problem from a serious breach, and the post doesn't answer either one. If you're deciding whether to save this, the useful thing to look for is the original log4j write-up, which should contain the packet captures or code paths behind the claim. Treat the post as a pointer to a possible story rather than a story itself. The specific technical details are the kind that can be tested, and until someone publishes that testing, the accusations of criminality and the "worst ever" framing outrun the evidence. Save it if you want to follow the original research, but read that research before forming a view.

Ask about this article