Alexa, Are You Listening?
The market for smart-home goods continues to grow, thanks in part to Amazon, which launched a dozen new hardware and software products in late 2018.[1] The commerce and cloud computing company is a leader in the smart-home industry, demonstrating a commitment to improving and streamlining menial tasks.[2] The brand announced new products that expand on Amazon’s popular Echo devices,[3] including an app-operated assistant, the Echo, and a voice-activated microwave.[4] Amazon’s Echo is a competitor among other industry leaders’ smart assistants: Apple’s Siri, Google’s Home, and Microsoft’s Cortana.[5]
Amazon’s voice-controlled devices can complete user-requested tasks, such as playing music, setting timers, turning on linked devices, and providing real-time information.[6] However, the device does much more than answer commands—it listens, learns, and saves data.[7] By using the device, consumers forfeit their privacy and confidentiality rights within their home for the convenience of a hands-free experience.[8]
The Echo, commonly referred to as “Alexa”, collects ambient sounds and active search requests in Amazon’s cloud data storage.[9] This information is known as voice data.[10] Voice data is similar but more valuable than typical text-based data promulgated by internet search engines.[11] Voice data saves a digital fingerprint of the speaker’s voice—including the tone, intensity, and intent—and other unique characteristics of the speaker’s identity and state of mind.[12] The digital fingerprint could expose intimate information about the user, like whether they are happy or sad, mourning or celebrating, tired, or awake. From that, voice data draws conclusions based on the specific intent of the user.[13] By comparison text-driven data is incapable of uncovering details of this nature; it is limited to key words, grammar and punctuation.[14] The potential future use of valuable voice data raise concerns about users’ rights to privacy and confidentiality.
Questions of privacy protection are increasingly relevant as voice-controlled devices expand beyond smart assistants and infiltrate more private spaces. The Supreme Court has acknowledged a First Amendment right to, and expectation of, privacy in specific scenarios.[15] Similarly, the U.S.A. FREEDOM Act reaffirmed a person’s right to privacy and security by limiting the National Security Agency’s power.[16] However, in Smith v. Maryland,[17] the Supreme Court held that the person who voluntarily gives information to a third-party forfeits a reasonable expectation of privacy.[18] This decision is most problematic in regards to voice-controlled devices.
Consumers consent to the ever-changing terms and conditions of voice-controlled devices simply by powering-on and using the device.[19]. In the terms and conditions, there is inevitably a clause stating the companies right to distribute data to third parties.[20] In Amazon’s terms of use, it states that a third party service is used with every search for the weather, music stations, auxiliary products, and other requests as needed.[21] However, in Jian Zhang v. Baidu.com, Inc.,[22] the court did recognize First Amendment protection for search engine results powered by computer algorithms. The court attributed protection for the search engines results because the output was a curation of content.[23] Applying this standard, the third party limitation may not extend to smart assistants if courts determine the devices output to be curated.
It is also suggested that the Fourth Amendment offers heightened privacy rights for voice-controlled devices and data.[24] Voice-controlled devices are at risk of being remotely controlled or accidentally engaged by erroneous ‘wake words’,[25] because the device is required to be constantly listening for its ‘wake word’.[26] This continuous engagement risks errant recordings—unauthorized by the device user—thereby triggering Fourth Amendment protections. First, in order for the Fourth Amendment to be relevant, there must be a reasonable expectation of privacy.[27] Even then, users will still need to overcome the third-party doctrine limitation.[28]
Smart assistants were heavily discussed during the investigation of the murder of Victor Collins.[29] Investigators attempted to collect the voice data from defendant James Bates’ Echo device. [30] Initially, Amazon agreed to hand over text-transcripts to the court, but refused to provide any voice data recorded by the device.[31] Sometime after, Bates granted authorities access to the voice data, prompting Amazon to relinquish copies of the recordings.[32]
While the investigation was ongoing, questions of how much voice data the Alexa recorded came into question.[33]. It was and remains unclear whether the device in question initiated errant or perpetual recordings during the period in question.[34]. However, Amazon ultimately avoided First and Fourth Amendment questions because it was given permission by the defendant to release its data.[35].
Voice-controlled devices and voice data risk an invasion of privacy in intimate and public spaces people occupy on a daily basis: rooms in the house, the workplace, even the car. The risk will continue to grow as the technology is integrated into products like children’s toys.[36] Amazon has argued that the commands and device responses constitute protected speech.[37] This argument is reinforced by the company’s unwillingness to hand over voice data in the Collins investigation. Other major players in the voice-controlled device market have and are likely to take a similar stance to retain customer trust. Consumers have and continue to fear government tracking and censoring.[38] As their convenience ushers them into the home, the threat of lost rights and unclear protections threatens the public perception of voice-controlled devices.
Footnotes
