Digital assistants such as Amazon's Alexa and Google's Assistant are still regularly being mistakenly activated when they hear what they think are their activation phrases. — dpa
Did someone say Alexa? For all their ability to tell us how tall Barack Obama is and report the weather, digital assistants such as Amazon's Alexa and Google's Assistant are still regularly being mistakenly activated when they hear what they think is their activation phrases.
In a test by a German consumer protection group, assistants were found to respond not only to terms that are similar to their activation words but also ones that are quite different.
Testers found that the microphone of Amazon's assistant Alexa was activated by the words "Alexandra" and "Gecko" (instead of "Echo"). They also found that the Amazon Echo speaker could even be accidentally activated by background noise.
Google's Assistant also responded to phrases that sound similar to its activation words, according to North Rhine-Westphalia Consumer Centre which conducted the test.
What's more, turning off the microphone didn't appear to stop data from being transferred between the app, the voice assistant and remote servers. This data is encrypted, meaning it's not possible for testers to say what it consists of.
However they did find that the overall data security levels of the voice assistants can be considered relatively high.
It was also found that data transfer was not exclusively linked to the voice assistants fulfilling their functions, which may be related to an advertising function.
Google integrates an ad server that delivers and measures the success of Internet advertising, while Amazon requires its users to consent to cookies and Internet ads when they're installing the app. — dpa
