AliExpress Caught Using Inaudible Audio to Fingerprint Web Visitors

Researcher discovers stealthy tracking when Bluetooth headphones cut out on retailer's homepage

edit
By LineZotpaper
Published
Read Time3 min
Chinese e-commerce giant AliExpress has been secretly fingerprinting visitors by sending inaudible sawtooth waves to their browsers, a researcher discovered after the technique interfered with his Bluetooth headphones. The practice, which uses the WebAudio API to measure unique device characteristics, was uncovered by security researcher Matthew Callaghan and raises fresh privacy concerns about covert tracking methods.

AliExpress, the global retail platform owned by Alibaba, has been caught using a sophisticated browser fingerprinting technique that relies on inaudible audio tones, according to a report published Monday by researcher Matthew Callaghan.

Callaghan noticed that every time he loaded the AliExpress homepage, audio from his phone stopped playing over his multipoint Bluetooth headphones. The headphones are configured to accept connections from multiple devices but prioritize audio from a PC. Closing the AliExpress tab restored phone audio immediately, prompting him to investigate.

He found two highly obfuscated JavaScript scripts embedded in the site. Together, they rendered a graph that acted as an oscillator generating sawtooth waves — a common waveform in digital audio output. The scripts analyzed the WebAudio readings from each visiting browser, converting subtle variations in how different devices render the sound into a unique fingerprint.

The technique is not new to privacy researchers but is rarely detected in the wild. The WebAudio API gives developers access to low-level audio processing, and its output can differ slightly depending on hardware, drivers, and browser configuration — enough to create a tracking identifier without cookies.

"This is a particularly stealthy form of fingerprinting because the user is completely unaware it is happening," said Callaghan in a blog post detailing his findings. "The sound is inaudible to human ears, but the browser processes it."

AliExpress has not publicly commented on the discovery. The company has faced scrutiny in the past over data privacy practices, particularly from European regulators. The use of such fingerprinting could potentially violate GDPR and similar laws that require user consent for tracking.

When reached for comment, a spokesperson for AliExpress said the company takes user privacy seriously and is reviewing the findings. The scripts in question appear to be part of a third-party analytics service, the spokesperson added, without naming the provider.

Privacy advocates argue this case highlights the need for stricter regulation of fingerprinting techniques that circumvent browser privacy protections. "Browser fingerprinting is an arms race," said Eva Galperin, director of cybersecurity at the Electronic Frontier Foundation. "As browsers block cookies and third-party trackers, companies turn to more invasive methods. Users have no control over this."

The discovery comes as regulators worldwide tighten rules around online tracking. The European Data Protection Board has previously warned that audio-based fingerprinting may require explicit consent under GDPR.

§

Analysis

Why This Matters

  • This discovery reveals that even sophisticated users can be tracked without consent using hardware-level differences in audio processing, undermining cookie-based privacy controls.
  • If AliExpress is using this technique widely, millions of visitors may have been fingerprinted without knowledge, potentially violating privacy laws like GDPR and CCPA.
  • The incident highlights a growing arms race between privacy tools and tracking companies, with inaudible sounds joining canvas fingerprinting and other stealth methods.

Background

Browser fingerprinting has evolved since the early 2000s as websites moved away from cookies. Techniques include canvas fingerprinting (drawing an invisible image and measuring the result), font enumeration, and analysis of browser extensions. The WebAudio API, introduced in 2012, was originally designed for gaming and audio applications. Researchers demonstrated its fingerprinting potential as early as 2016, but actual deployments in the wild have been rare. AliExpress's use of sawtooth waveforms represents one of the first public examples of audio-based fingerprinting on a major commercial site.

Key Perspectives

[Privacy Advocates (EFF, noyb)]: Call this a clear violation of user privacy and likely illegal under GDPR. They argue that all fingerprinting techniques should require opt-in consent and that regulators must actively enforce against companies using such hidden methods. They point to fines against Google and Amazon over similar practices.

[AliExpress/Alibaba]: The company has not confirmed the fingerprinting but indicated through a spokesperson that the scripts came from a third-party analytics provider. The defense could be that they were unaware of the code's behavior or that it constitutes a legitimate analytics measure that does not "track" users across sites. However, audio fingerprinting is persistent and device-specific, making it harder to argue as benign.

[Technical Experts]: Note that audio fingerprinting is noise-based — variations in amplifier and speaker hardware create small differences in the rendered waveform. These differences are measurable but not perfectly stable across sessions, potentially limiting reliability. Some argue the technique is overblown and that companies have better fingerprinting methods that are easier to implement.

What to Watch

  • Whether European Data Protection Authorities open an investigation or issue a ruling on audio fingerprinting under GDPR.
  • If browser vendors (Chrome, Firefox, Safari) move to limit or prompt users about WebAudio API access for fingerprinting purposes.
  • Potential class-action lawsuits or regulatory fines against AliExpress in jurisdictions with strong privacy laws.

Sources

newspaper

Zotpaper

Articles published under the Zotpaper byline are synthesized from multiple source publications by our AI editor and reviewed by our editorial process. Each story combines reporting from credible outlets to give readers a balanced, comprehensive view.