Alexa Smart Speaker Investigation Earns Privacy Researchers Best Paper Award

The ProperData Center continues to research various aspects of smart speakers, and more generally privacy aspects of emerging ways we interact with smart devices.

Nov. 20, 2023 - A multi-university research team from ProperData Center won the Best Paper Award for its study showing that Amazon collects and uses smart speaker voice data to target advertisements. The award was presented at the Association for Computing Machinery’s 2023 Internet Measurement Conference in Montreal.

The paper, “Tracking, Profiling, and Ad Targeting in the Alexa Echo Smart Speaker Ecosystem,” posted in 2022, garnered attention from the media and Amazon, which subsequently made policy updates.

The research team included lead author Umar Iqbal (formerly a postdoc at University of Washington and ProperData alum, now faculty at Washington University, St. Louis) and Franziska Roesner (also at UW); Pouneh Nikkhah Bahrami and Zubair Shafiq (UC Davis); Athina Markopoulou, Rahmadi Trimananda and Hao Cui (UC Irvine), Alexander Gamero-Garrido (formerly a postdoc at Northeastern, now faculty at UC Davis); Daniel Dubois and David Choffnes (Northeastern University).

“This work is one of the success stories of the ProperData project, with a collaboration across our institutions,” said Markopoulou, UCI professor of electrical engineering and computer science and one of the authors. “The paper showed that Amazon Echo smart speakers, and third-party skills running on them, actually use voice interaction data to target people with advertising, despite previous disclosures on this matter.”

The paper had impact: it generated public debate, was followed by changes in Amazon's privacy policy statements and was also presented at PrivacyCon 2022– a flagship event hosted by the Federal Trade Commission.

“ProperData continues to research various aspects of smart speakers, and more generally privacy aspects of emerging ways we interact with smart devices, for example through modalities such as voice, powered by large language models,” said Markopoulou.

Markopoulou directs the ProperData Center, a five-year, $10 million NSF SATC Frontiers project that addresses the urgent need for protection of personal data flow on the internet. Its  research seeks to provide more transparency and control over personal data, using an approach that combines technology and policy.

Click here for more information on the paper.

– Lori Brandt