Short answers, pulled from the story.
Speaker recognition is the identification of a person from the acoustic characteristics of their voice, answering the question "who is speaking" rather than "what is being said." Each system has two phases: enrollment, where a voice print or template is created, and verification or identification, where a new speech sample is compared against stored prints.
Speaker verification is a 1:1 match that checks a claimed identity against a single stored template, while speaker identification is a 1:N match that compares an unknown voice against multiple templates to find the best match. Verification requires the user's cooperation and knowledge; identification can be performed covertly.
One of the earliest commercial deployments was in Worlds of Wonder's Julie doll in 1987, marketed with the tagline "Finally, the doll that understands you." Children could train the doll to respond to their specific voice. An international patent for speaker recognition was filed even earlier, in 1983, by Michele Cavazza and Alberto Ciaramella at CSELT in Italy.
In 2013, Barclays Wealth became the first financial services firm to deploy voice biometrics as the primary means of identifying customers to its call centers. The system verified caller identity within 30 seconds of normal conversation, and 93% of customers rated it 9 out of 10 for speed, ease of use, and security.
In 2023, both Vice News and The Guardian independently demonstrated they could defeat standard financial speaker-authentication systems using AI-generated voices cloned from approximately five minutes of a target's voice samples. The two demonstrations were separate and reached the same conclusion.
The General Data Protection Regulation in the European Union and the California Consumer Privacy Act in the United States have both prompted significant discussion about workplace deployment of speaker recognition. In September 2019, Irish speech technology company Soapbox Labs specifically warned about the legal implications of using such systems in employment settings.