A massive 73% of people don’t trust AI voice technology to make simple calls – although trust does grow as they use the technology.
Riley Panko, senior content developer and marketer at Clutch discusses the findings from recent research.
Nearly three-quarters of people (73%) say they are somewhat or very unlikely to trust a tool such as Google Duplex to correctly make simple calls for them. Duplex is a tool within Google Assistant that can call restaurants and book reservations using an AI-powered voice. Duplex caused controversy when it was unveiled in spring 2018 due to its lifelike realism.
This data comes from Clutch, the leading B2B ratings and reviews firm, and Ciklum, a global digital solutions company. Clutch and Ciklum surveyed 501 people who called businesses at least three times in the past six months to understand their comfort level with conversational AI tools.
Experts mentioned that people often feel hesitant toward new technologies that later become accepted, such as ATMs or self-checkout lines.
“To varying degrees, we’re all vulnerable to different forms of this anxiety when it comes to adjusting to new ideas or technological advances,” says Ivan Kotiuchyi, research engineer at Ciklum.
AI voice technology presents new obstacles for winning consumer confidence due to its lack of monitoring, though.
Yet, experts say that as a tool like Duplex becomes more commonplace and demonstrates its success, people will grow to trust it.
More than eith in 10 people (81%) want AI voice technology such as Duplex to declare itself as a robot before proceeding with a call.
This desire for identification shows that people may be wary of the ways in which AI voices can be used for manipulation.
Criminals may be able to use conversational AI to automate increasingly realistic scam calls.
Overall, the survey indicates that people are presently uncomfortable with conversational AI but will likely embrace it as it grows in popularity. People should be cautious of scams that exploit the technology behind Google Duplex for malicious purposes, though.