Former AWS security engineer says Amazon's new Bee AI wearables leave bystanders out of the consent loop

Amazon’s new Bee AI devices put their consent safeguards in front of the wrong person, according to Jeff Williams, a former senior principal security engineer at Amazon Web Services, GeekWire reported on Friday. In a LinkedIn article, Williams wrote that the people being recorded have no practical way to say no.

ADVERTISEMENT

ARQS OFFERIs your business AI-ready?Get a focused AI Readiness Audit with practical next steps.$99 one-timeStart the audit →

Amazon unveiled the devices on Thursday, its first new Bee hardware since it acquired the San Francisco startup last year. The $74.99 Bee Frontier is worn on the wrist or clipped to clothing, and the $99.99 Bee Eclipse attaches to the back of a phone. Both record conversations and turn them into summaries and to-do lists, with connections to Amazon.com and Alexa+ for paid subscribers.

Amazon says users are prompted to get permission from participants when they start a recording and that an indicator light shows while the devices are in use. Williams says the prompt appears only on the wearer’s phone, nothing enforces it and the light is easy to hide. About a dozen U.S. states, including Washington and California, generally require everyone’s consent before a private conversation is recorded. He credited Bee Private Compute, which processes audio in an isolated, encrypted environment built on AWS Nitro, but said it protects the wearer rather than the people nearby.

Williams, now chief information security officer at Sigma360, suggested that Bee leave other people’s voices out of transcripts by default unless they agree. Amazon did not respond directly to the critique but pointed to an automatic shutoff when the owner’s voice is not detected for 5, 15, 30 or 60 minutes, in-app guidance on responsible use and reminders to tell people nearby and ask permission.