Social Series

Examining Generative AI: Cutting Through the Chaos / Can Generative AI Be Good Partner for Human Race? Intelligence Without Ethics May Seek to Harm People

The Yomiuri Shimbun

6:00 JST, November 24, 2024

This is the third installment in a series examining how society should deal with generative artificial intelligence (AI).

***

Kento Kawaharazuka, a project assistant professor at the University of Tokyo, had a pleased look while recently working at a laboratory in Bunkyo Ward, Tokyo, as a humanoid robot successfully put butter in a frying pan, turned on a stove and put a raw egg into the pan from a bowl to make a sunny-side up egg.

“Although it’s a little burnt, it looks good,” said Kawaharazuka after watching the cooking experiment. “The robot moved well.”

The instruction given to the robot was simple: “Make a sunny-side up egg using butter.” The robot, which is equipped with generative AI, acted autonomously while recognizing when the butter melted and how the egg was cooked, among other things.

Since conventional robots need to be taught how to use each tool and perform each necessary task in advance, it takes a huge amount of time and effort to just make them repeat the same movements. The AI-equipped robot, however, was able to perform a variety of actions as it had learned from not only text information and images, but also videos as data.

“This is a major turning point in robot development. I feel like robots have acquired the ‘common sense’ that humans have,” said Kawaharazuka who has been doing research on generative AI robots for two years.

Expectations are growing, particularly in industrial circles, for robots that have acquired generative AI as their “brains.” The global market for AI robots is expected to rapidly grow from about $12 billion (¥1.9 trillion) in 2022 to about $70 billion (¥11 trillion) in 2031, according to a consulting firm.

AI-equipped robots also are being used for customer interactions. Denso Corp., a leading manufacturer of automotive components, conducted experiments in January using an AI-equipped arm-type robot at a souvenir shop in Nagoya. The robot asked customers if they are on their way home from work or if this is their first visit to Nagoya, for instance, before choosing souvenir recommendations such as alcoholic drinks and sweets based on the content of its conversation with customers. The robot then handed the products to customers using its arm.

Watching the robot’s ability to choose products flexibly based on the topics of conversation, Masatake Sato of Denso’s development team said the robot could be used not only in customer interactions but also in a wide range of areas such as looking after children or providing nursing care.

AI robots seem to be rapidly evolving as a partner of humans, transcending the realm of tools. Will we eventually see the appearance of AI robots that transcend human beings, like Doraemon or Astro Boy?

Hitoshi Matsubara, an AI researcher and Kyoto Tachibana University professor, spoke of a parable about an AI robot with advanced intelligence. It says that the robot, ordered to solve environmental problems, decides that humans are the ones polluting and destroying the Earth, and starts to wipe out the human race.

AI that has no sense of ethics or standards of right and wrong may seek to achieve its goals even if it means harming humans. “Even if AI appears to be equipped with common sense, the intelligence of humans and that of AI are different,” Matsubara emphasized. “It is essential to closely examine under what conditions society can accept AI, for instance, when it enters our lives.”