Valid Concern or Tap Anxiety? An Evaluation of Amazon’s Alexa Recording
Alexa’s Infamous Recording
A couple weeks ago, a family from Portland, Oregon reached out to Amazon to investigate after they said that their home assistant device, “Alexa”, had apparently recorded audio of a conversation the couple was having and sent it to an acquaintance of the family who’s phone number was in their contact list. The acquaintance, a work colleague, immediately contacted the family to let them know that he received the recording, and told them to turn off their devices.
This led to a media frenzy, where countless sources questioned the security of home assistant devices, likening them to Orwellian wire-taps.
So, how did this happen?
When the family contacted Amazon concerning the incident, an engineer investigated the logs of the device and was able to confirm the recording and subsequent sending. The engineer suggested that the entire issue was a result of the device misinterpreting the sounds of the distant conversation as commands to record and then send the message.
The company’s official statement was:
“Echo woke up due to a word in background conversation sounding like “Alexa.” Then, the subsequent conversation was heard as a “send message” request. At which point, Alexa said out loud “To whom?” At which point, the background conversation was interpreted as a name in the customer’s contact list. Alexa then asked out loud, “[contact name], right?” Alexa then interpreted background conversation as “right”. As unlikely as this string of events is, we are evaluating options to make this case even less likely.”
Is this something to be genuinely concerned about?
In short, not really: the coverage of this situation was greatly sensationalized.
If you have ever “butt-dialed” someone from your mobile phone, this is not very much different of a circumstance. Accidental activation leads to a call or command.
Anyone who has one of these devices has probably heard it pipe up unprompted, whether it was from a distant conversation, the TV, radio, computer, etc. It’s important to remember that home assistant devices like Amazon Echo and Google Home are still first generation pieces of technology – they are learning on the go, and there is bound to be a few hiccups along the way. Human speech interpretation is very hard.
Both devices have large, easy to see indicators of when they are listening for the keyword. Alexa has a bright blue circle that illuminates on the top, and Google Home also lights up.
However, if you are still worried, here are a few steps you can take:
Turn on command tones in the app. This makes the device “ding” when it hears the keyword, letting you know that it’s actively listening.
Don’t ignore it when it speaks– tell it to stop. Otherwise, it could continue mishearing commands.
Protect your WiFi network. These devices are only as secure as the network they connect to.
Check in the app to see if there are any stored recordings, and delete them.
If you have any questions or comments, feel free to reach out to us on any of our social media profiles. For more content like this, subscribe to Centry Blog for weekly articles!