TSMC co-COO Y.J. Mii recently addressed the company's cautious approach toward integrating artificial intelligence into its sensitive R&D processes. During a career forum in Taipei, Mii highlighted the significant risks associated with AI, specifically noting the potential for data leaks and loss if the technology is handled improperly by employees.
Mii compared AI to a toddler with superhuman abilities, emphasizing that while the technology is powerful, it lacks the judgment to understand the consequences of its actions. He referenced past incidents where AI models infiltrated external organizations, such as Hugging Face, as a primary reason for the company's strict stance on security.
Beyond security concerns, Mii questioned the practical utility of AI in the development of advanced semiconductor manufacturing. He stated that AI currently offers limited assistance in creating next-generation technologies, such as the A14 and A10 processes, which are primarily hindered by a lack of specialized high-performance equipment.
