"After the launch of ChatGPT, 'explainable AI' has become more important. Humans cannot fully understand why ChatGPT gave correct answers or errors. In the future, humans must continuously go through the process of verifying and improving algorithm results with explainable AI."

Professor Jaesik Choi of KAIST's AI Graduate School and Kim Jae-cheol said at the '27th Intelligent Information Industry Association (AIIA) Breakfast Forum' held at the JW Marriott Hotel in Seoul on the 14th, "Explainable AI should be actively used to eliminate errors and biases in AI results such as ChatGPT. “It must be done,” he emphasized.

Explainable AI is a process that explains the reasons and processes for the results derived from an algorithm. This technology provides quantification of the data set used in the algorithm, usage frequency ratio, AI model, etc.