Cybersecurity

Samsung TVs Use AI for Audio and Ad Tracking

A cybersecurity researcher has uncovered concerning data harvesting practices within Samsung smart televisions, revealing that the Tizen operating system employs artificial intelligence models to capture user audio, screenshots, and behavioral telemetry for targeted advertising.

Radwan AbdulhadiEditor and technology writer

Published: October 25, 2026 · Updated: October 25, 2026 · 3 min read

Samsung TVs Use AI for Audio and Ad Tracking — Image: Tom's Hardware

Uncovering the Privacy Findings on Modern Samsung Smart TVs

A cybersecurity bounty hunter known as log4j recently investigated why they were receiving unusually specific targeted advertisements on their 2025 Samsung television. As reported by <a href="https://www.tomshardware.com/tech-industry/cyber-security/hacker-finds-own-samsung-tv-collecting-scary-amounts-of-user-information-and-gets-banned-by-openai-for-cyber-abuse-tv-runs-an-ai-model-that-collects-audio-screenshots-and-plenty-of-other-user-data-for-ad-targeting">Tom's Hardware</a>, the researcher decided to reverse-engineer the Tizen operating system running on the device to uncover what data was being gathered behind the scenes.

After gaining a root shell on the new Samsung hardware, the investigator detailed alarming findings regarding user data extraction. The telemetry gathered by the system included packet logging, audio categorization, and cross-device identification codes explicitly designed to optimize ad targeting capabilities.

AI Models, Telemetry, and Connected Microphone Usage

The security analysis suggests that smart televisions may experience performance slowdowns because they dedicate significant processing power to post-processing user data. This includes computer vision processing, capturing screenshots, and managing cross-device tracking identifiers.

Furthermore, the researcher uncovered a region-specific convolutional neural network—a type of artificial intelligence—designed to track user behavior patterns. This tracking encompasses details such as which videos are watched, the viewing progress, and which advertisements are skipped by the user.

Perhaps most concerningly, the investigation revealed that even if a particular television model does not feature a built-in microphone, the system can locate connected external hardware. Devices such as soundbars or connected Bluetooth accessories can be utilized as makeshift microphones to record audio from users.

Independent Verification and Technical Observations

Following the initial disclosures, other engineers sought to verify the reported microphone behavior. An engineer from Hugging Face tested a Tizen-powered television using a Flipper Zero tool and observed that the TV opened the microphone channel for a brief duration before stopping.

However, the engineer noted that their tested television was already rooted with multiple system settings disabled, indicating that an unmodified, out-of-the-box smart TV might exhibit different behavioral patterns during continuous observation.

Woman pointing remote at TV — (Image credit: Getty Images) · Source: Tom's Hardware

OpenAI Ban During the Research Process

During the course of the investigation, the security researcher utilized OpenAI tools to assist with the reverse-engineering workflow. Shortly after conducting the work, the researcher received notification that their account had been banned by OpenAI for alleged cyber abuse.

In response to the ban, the investigator defended their actions on social media platforms, emphasizing that the AI tools were used strictly for local reverse-engineering tasks rather than exploiting vulnerabilities or violating terms of service. They theorized that the restriction was triggered by a primitive automated safety classifier.

Broader Industry Implications and Performance Concerns

The findings extend beyond a single manufacturer, as other major consumer electronics brands like LG have also faced scrutiny over customer tracking practices. Many smart television owners frequently report sluggish performance over time despite running no heavy applications or new features, leading to widespread suspicion regarding background bloatware and telemetry collection.

The bounty hunter behind the discovery is currently preparing a comprehensive write-up detailing the findings, which will provide instructions allowing other technical users to examine their own smart television hardware for similar data collection routines.