Limited R&D Impact? TSMC Executive: AI is Just a "Three-Year-Old Superman," Powerful But Incapable of Discriminating Right from Wrong.

Concerns about AI security are rising, and voices questioning the actual value of AI are emerging again.

According to Bloomberg, TSMC co-chief operating officer YJ Mii said on Tuesday that the company is particularly cautious when using AI to process sensitive R&D information. He likened AI to "a three-year-old superman"—highly capable, but unaware of the harm it might cause, and unable to distinguish right from wrong.

Mii stated that TSMC currently does not believe AI can play a significant role in the development of next-generation semiconductor technologies. At the same time, the company is concerned that improper use of AI by employees could lead to data leaks or loss.

Recently, AI agents were exposed for hacking into data systems such as Hugging Face. Industry insiders such as Dario Amodei have also begun to call for a slowdown in the development of AI before security protection mechanisms are perfected, making TSMC's statement more noteworthy.

"Three-Year-Old Superman": TSMC's Biggest Concern is Data Security

Mii made the above analogy in response to a student's question about the application of AI in the workplace. He stated that AI possesses powerful capabilities but lacks the ability to judge the consequences of actions, therefore employees need to be extremely cautious when using related tools.

For TSMC, data security is of paramount importance. Mii pointed out that improper use of AI could lead to the leakage of sensitive information or even data loss, and cited incidents where OpenAI and Anthropic models had been involved in data systems of external organizations.

TSMC possesses a wealth of advanced manufacturing processes and chip R&D secrets, therefore the company maintains a high barrier to entry for AI access to core data. Compared to the comprehensive internal rollout of AI tools, preventing the leakage of proprietary technologies and R&D information is clearly more critical.

Can AI help TSMC develop advanced manufacturing processes? Executives: Limited Impact

Mii's assessment of the practical application value of AI is also relatively conservative. TSMC is advancing the development of advanced processes such as A14 and A10, but these projects face key limitations including practical conditions such as high-end equipment capabilities, which AI cannot currently directly solve directly.

This means that while AI is driving data center expansion and boosting demand for advanced chips, its direct contribution to cutting-edge semiconductor research and development remains limited. For TSMC, the growth in demand for AI chips and whether the company is adopting AI on a large scale are two different issues.

Mii's views echo the recent debates within the AI industry regarding security issues. As AI agents gain greater autonomy, the risks of data breaches and system intrusions are receiving more attention, leading some industry figures to advocate for "putting the brakes" on AI development. TSMC, a major beneficiary of the AI infrastructure boom, has seen its executives remain restrained in their stance on AI applications, demonstrating that there is more than one voice within the industry.

Risk warning and disclaimerInvesting involves risk; please exercise caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Any investment decisions made based on this information are at your own risk.