Bloomberg has spoken with Amazon contractors who say they listen to 1,000 Alexa audio clips per shift, on average, to review Alexa's response rates and identify glitches.

Amazon's AI audio review process appears more thorough and involved than those run by Apple or Google. Some contractors, for example, described how they were tasked with categorizing topics users ask Alexa about:

One worker in Boston said he mined accumulated voice data for specific utterances such as “Taylor Swift” and annotated them to indicate the searcher meant the musical artist.

Like other tech companies, Amazon's privacy policy at the time of Bloomberg's initial story -- April, 2019 -- did not explicitly state that humans review audio clips recorded by Alexa.

A job opening Bloomberg found at the time described the reviewer's job, suggesting that the technology behind Alexa's responses still relies heavily on human curation:

A recent Amazon job posting, seeking a quality assurance manager for Alexa Data Services in Bucharest, describes the role humans play: “Every day she [Alexa] listens to thousands of people talking to her about different topics and different languages, and she needs our help to make sense of it all.” The want ad continues: “This is big data handling like you’ve never seen it. We’re creating, labeling, curating and analyzing vast quantities of speech on a daily basis.”