[Opinion] Industrial AX: Where to Start Properly?

Industrial AI transformation, or AX, is conceptualized as bestowing 'eyes, ears, a nervous system, and a brain' upon industrial assets. Sensors and hardware act as the eyes and ears, detecting the physical world, while data transmission systems form the nervous system. Engineering knowledge interprets these signals, and AI serves as the brain, predicting future events and guiding necessary actions.

As the government designates Physical AI as a core component of its 'Grand Leap Forward 3 Mega Projects,' there's an accelerated push to link AI with real-world industrial data. Efforts are underway across manufacturing, mobility, energy, and plant sectors to integrate equipment data and translate AI analysis into operational decisions.

However, simply applying AI models to existing facilities does not automatically achieve industrial AX. The article questions the focus on AI model performance, asking, 'How accurately is the AI seeing the current reality?' It suggests that a more fundamental question is whether the AI is truly connected to and accurately perceiving the actual state of industrial operations. A comprehensive approach is needed, integrating various AI models and leveraging existing logs and data, rather than solely focusing on the AI's predictive power in isolation.