TSMC Executive Compares AI to a “Three-Year-Old Superman” as Chip Industry Weighs Its Future
A senior executive at Taiwan Semiconductor Manufacturing Company has offered a striking comparison for the current state of artificial intelligence, describing AI as a “three-year-old Superman” — incredibly powerful, but still lacking the judgment and maturity needed to understand the consequences of its actions.
Y. J. Mii, co-chief operating officer at TSMC, made the remarks while speaking to students at National Taiwan University during an event organized by United Daily News, the TSMC Education Foundation, and the university. His comments come at a time when the global technology industry is increasingly debating how far AI should be allowed to go, especially in sensitive areas such as semiconductor design, software development, and corporate decision-making.
According to Mii, AI already plays an important role inside TSMC, particularly in coding-related tasks. He explained that programming is one of the areas where current AI tools can be highly effective because the work is relatively structured and easier for AI systems to assist with. In that sense, AI can improve efficiency, reduce repetitive work, and support engineers in their daily tasks.
Mii also noted that artificial intelligence can be useful in chip design and electronic design automation, commonly known as EDA. This area involves using software tools to design, verify, and optimize semiconductor chips. Because chip design often relies on defined inputs and outputs, AI can help engineers work through complex design challenges more efficiently.
The semiconductor industry is already seeing growing interest in AI-assisted design tools. These systems can help optimize layouts, test design possibilities, and support the development of chips containing millions or even billions of transistors. For certain tasks, AI can process data and explore options faster than humans, making it a valuable tool in modern chip engineering.
However, Mii made it clear that AI has limits, especially when it comes to the most advanced semiconductor manufacturing technologies.
TSMC is currently working on next-generation manufacturing processes, including 1.4-nanometer technology, known as A14, as well as technologies beyond that. These advanced nodes are expected to play a critical role in the future of high-performance computing, artificial intelligence processors, mobile chips, and other cutting-edge applications.
But Mii believes AI is not yet suitable for designing leading-edge chip manufacturing technologies from the ground up. The reason, he explained, is that future process technologies involve too many unknown variables. AI depends heavily on existing data, and when a technology has not yet been developed, there is limited real-world information for AI systems to learn from.
In other words, AI can be powerful when it works within known boundaries, but it struggles when the task requires discovering something entirely new. If current equipment, materials, or manufacturing methods cannot achieve a specific goal, AI cannot magically remove those physical limitations.
This is why Mii emphasized that human expertise remains essential. In advanced semiconductor research, engineers often face problems that have no historical data, no proven solutions, and no clear roadmap. These challenges require scientific judgment, creativity, experimentation, and experience — qualities that AI cannot fully replace.
His “three-year-old Superman” analogy captures this concern clearly. AI may have extraordinary capabilities, but without mature understanding, it can also create risks. That is why Mii believes humans must remain responsible for making final decisions when using AI, particularly in mission-critical industries such as semiconductor manufacturing.
He also pointed to another important concern: information security. At the corporate level, companies must be careful about what they share with AI systems because of the potential risk of sensitive data leaks. For a company like TSMC, which operates at the center of the global chip supply chain, protecting confidential manufacturing knowledge is a top priority.
Beyond AI, Mii also discussed TSMC’s global expansion. The company has been building and planning advanced semiconductor facilities outside Taiwan as demand for chips continues to grow worldwide. However, expanding internationally comes with major challenges.
One of the biggest hurdles, according to Mii, is finding experienced senior managers capable of leading global teams. Building chip fabrication plants is not only about technology and investment. It also requires strong leadership, deep operational knowledge, cultural understanding, and the ability to manage complex teams across different countries.
TSMC’s expansion strategy reflects the growing importance of semiconductors in the global economy. Chips are now essential for artificial intelligence, smartphones, data centers, vehicles, industrial systems, and defense technologies. As countries seek stronger domestic chip supply chains, companies like TSMC are under pressure to scale production while maintaining their industry-leading standards.
Mii’s comments highlight a balanced view of artificial intelligence in the semiconductor industry. AI is already transforming parts of chip design and software development, and its role will likely continue to grow. But when it comes to pioneering new manufacturing technologies, human engineers remain at the center of innovation.
For TSMC, the future of chipmaking will not be driven by AI alone. Instead, it will depend on the careful combination of artificial intelligence, human expertise, advanced materials, precision equipment, and strong global leadership. AI may be powerful, but in the race to build the next generation of semiconductors, judgment and experience still matter most.
