World News

South Korea telecom firm aims to launch AI kiosks for customer service

By Nam Gyeong-sik & Kim Tae-gyu, UPI News Korea   |   April 19, 2022 at 9:48 AM
KT Senior Vice President Choi Joon-ki (R) holds a memorandum of understanding with Deep Brain AI CEO Eric Jang at the Deep Brain AI headquarters. Photo courtesy of KT

SEOUL, April 19 (UPI) -- South Korea's largest telecom company has announced plans to launch kiosks that can communicate with humans this year.

KT has signed a memorandum of understanding with Deep Brain AI, a company specializing in visual artificial intelligence technologies, KT announced Sunday.

Advertising
Advertising

"Thus far, kiosks have helped people by carrying out preset functions. But our AI will enable kiosks to talk with people on a real-time basis," a KT spokesman who agreed to speak on the condition of anonymity told UPI News Korea.

"In addition, the AI [entity] will be able to provide 24-hour non-face-to-face services through other platforms like robots, home appliances, mobile apps and cars," he said.

The company plans to name the service "Giga Gini AI Human."

"We are cooperating with universities and research institutes to keep developing a mega-scale AI that has conversational skills as natural as that of a human being," the spokesman said.

Analysts say KT is trying to pass the Turing test.

"There are controversies, but most experts believe that AI still hasn't passed the Turing test yet," South Korean business tracker Leaders Index CEO Park Ju-gun said in a telephone interview.

"Some question the viability of the Turing test and the value of pursuing it. But the test is a kind of holy grail for many AI researchers. Many companies want to pass the Turing test for the first time in the world without any doubts," he said.

Named for British computer scientist Alan Turing, the test judges whether a computer can "think" like a human being.

In the test, human evaluators judge natural language conversations between a human and a machine. To pass the Turing test, the evaluator should not be able to tell the difference between a machine and a human.