[News Terms] 'AGI' That Reasons and Acts Like a Human

'AGI (Artificial General Intelligence)' refers to the intelligence of machines capable of handling data such as language, images, videos, and speech with intelligence close to that of a typical human. In Korea, the terms 'Beomyong In-gong Jineung' or 'In-gong Ilban Jineung' are used.

It means that AI (Artificial Intelligence) has reached a stage where it can think and grow on its own without human intervention. It also refers to AI or research that has the ability to think, learn, and create in all given situations, not just specific problems, and is one of the ultimate goals of AI research.

On the 16th (local time), Sam Altman, CEO of OpenAI, which launched the generative AI ChatGPT, appeared as a witness at a U.S. Senate hearing and urged regulation of AI, stating, "If this technology goes wrong, it can cause big problems."
[Photo by AP/ Yonhap News]

Because it can think and carry out tasks on its own without human commands, it is also called 'Strong AI' or 'Complete AI.' AI that implements software to solve specific problems, such as AlphaGo or Deep Blue, which play Go or chess by inputting basic data, algorithms, and rules of the program, is called 'Weak AI' or 'Applied AI.'

According to a report by The New York Times (NYT) on the 16th (local time), researchers affiliated with Microsoft (MS) claimed that AI has approached the AGI stage through the paper 'The Spark of AGI' published last March.

In experiments on AI's problem-solving ability, researchers asked how to stably stack a book, nine eggs, a laptop, a bottle, and nails. The AI responded as if it had cognitive ability, saying, "Arrange the eggs in three rows on top of the book with spaces in between. If you place the laptop on the eggs, it will fit perfectly at the boundary between the book and the eggs." Some experts, including Marton Szep, a professor at Carnegie Mellon University, dismissed this as "corporate promotion disguised as a paper," but it is true that an important milestone in AI research has been reached.

In this regard, Sam Altman, CEO of OpenAI, which released the generative AI ChatGPT, appeared as a witness at a U.S. Senate hearing on AI regulations on the same day and urged regulation of AI, saying, "If this technology goes wrong, it can cause big problems."