Want CNET to notify you of price drops and the latest stories?

Phone scammer or trustworthy solicitor? Software calls it

Japanese researchers aim to protect potential victims of phone scams by identifying "overtrust" in their voices when they interact with the bad guys.

Leslie Katz Former Culture Editor
Leslie Katz led a team that explored the intersection of tech and culture, plus all manner of awe-inspiring science, from space to AI and archaeology. When she's not smithing words, she's probably playing online word games, tending to her garden or referring to herself in the third person.
  • Third place film critic, 2021 LA Press Club National Arts & Entertainment Journalism Awards
Leslie Katz
2 min read
The new system aims to detect phone scams and alert the intended victim, as well as the intended victim's family or another objective third party Fujitsu

Should you trust that official-sounding guy on the phone trying to talk you into transferring cash from your account to his? Well, we're going to go out on a limb here and say no, but in case that's not so obvious, new voice analysis technology out of Japan promises to help spot the scammers for you.

Nagoya University and Fujitsu have created software they say can automatically identify situations in which one party might "overtrust" the other. It does so by detecting changes in voice pitch and volume level that can occur under psychological duress.

By combining this technology with keywords such as "indebtedness" or "compensation" that are characteristic to a specific type of remittance-soliciting phone-phishing scam called furikomesagi, the researchers have developed a setup now being tested in collaboration with the National Police Agency of Japan and the Bank of Nagoya.

Furikomesagi ploys, which the researchers dub a "serious social problem in Japan," typically target the elderly over the phone.

Perpetrators often pose as an acquaintance of the intended victim or a person in a position of authority, such as a police officer or lawyer, to convey distressing information relating to a scandal or crime supposedly committed by an acquaintance of the victim. Surprise, surprise, it's all meant to deceive people into wiring large amounts of money.

The new system aims to protect victims of such schemes by alerting them, as well as family members or other objective third parties, that a ruse is afoot. Tests of the technology identified remittance fraud with more than 90 percent accuracy, the researchers said.

While speech recognition systems have previously employed keyword-detecting technology, the new system goes a step further by also taking into consideration the psychological stress that can occur when someone is receiving distressing information that catches them by surprise, confuses them, or impairs their judgment.

Meanwhile, the scam-stopping setup joins a growing list of technologies that purport to understand the subtleties of human character and communication in voice, writing, and even photos. Among these are Exaudios' system, which examines phone callers' intonations to tell if they are happy, sad, or angry; and ToneCheck, a "spell-checker" for sentiment in e-mail.

Israeli researchers have developed one tool that spots words, phrases, and even metaphors indicating depression in online text, and another that purports to spot sarcasm. And yet another tool examines images of faces for traits such as competence, trustworthiness, meanness, dominance, and extroversion.