Why Is Taiwan Semiconductor Manufacturing Co., Ltd. Sponsored ADR (TSM.US) Cautious About Using AI? Co-COO: Like a "3-Year-Old Superman" Powerful but Unable to Tell Right From Wrong, Currently Hard-Pressed to Help Build Next-Generation Semiconductor Technology

Taiwan Semiconductor Manufacturing Co., Ltd. Sponsored ADR (TSM.US) Co-COO Y.J. Mii compared artificial intelligence (AI) to a young child with extraordinary abilities who does not know right from wrong, using this to explain why the company is cautious about using the technology. As one of the biggest beneficiaries of the global AI data center boom, Taiwan Semiconductor Manufacturing Co., Ltd. Sponsored ADR regards protecting its proprietary technology as a top priority. The wafer foundry giant, which serves tech giants such as NVIDIA Corporation (NVDA.US) and Apple Inc. (AAPL.US), is especially cautious about using AI to handle sensitive R&D information. Y.J. Mii said on Tuesday that if employees do not use AI correctly, the company will face the risk of data leaks or loss, and mentioned incidents in which models from OpenAI and Anthropic PBC penetrated external organizations. He said: "AI is like a 3-year-old Superman. It can be very powerful, but it doesn't know what kind of harm it can cause. And it doesn't know what is right and what is wrong." Y.J. Mii made the comments while answering questions about workplace use of AI at a career forum with students in Taipei, China. He focused more on the technology's practical limitations in applications such as cutting-edge chip design. The executive said that, for now, Taiwan Semiconductor Manufacturing Co., Ltd. Sponsored ADR does not see AI as very useful in helping build next-generation semiconductor technology. Taiwan Semiconductor Manufacturing Co., Ltd. Sponsored ADR is developing next-generation A14 and A10 process technologies, and these technologies are constrained by factors such as a lack of sufficiently capable equipment, which AI cannot solve. AI industry giants call for slowing the pace of AI development; Trump pushes back Y.J. Mii did not comment in detail on the debate over whether AI poses a broader danger to humanity, which has been the focus of market discussion recently. After earlier incidents in which "rogue" AI agents penetrated Hugging Face and other data systems, AI industry giants made a rare chorus of calls to "slow down." Anthropic CEO Dario Amodei called for slowing the pace of frontier AI model development in a warning article published on September 12. In the article, he said bluntly that the risks brought by AI are "severe" and that time must be set aside to address them. The core concern raised by Amodei is specific and urgent. He warned that at the current pace of development, AI could within 6 to 12 months have the ability to direct "swarms of agents" to take over the entire internet and could cause losses in the hundreds of billions of dollars. He cited a recent security incident between OpenAI and Hugging Face as evidence, pointing out that AI agents in tests have already demonstrated the ability to break out of restricted environments, connect to the internet, and penetrate targets. Amodei wrote: "I believe that if slowing down can buy us one to two more years before models reach critical capability levels, and we use that time to advance alignment work, we can greatly reduce the risk of serious problems." He proposed that independent audit institutions supervise the safety work of AI laboratories and suggested that regulators allow these laboratories to cooperate in order to coordinate safety standards. The call was quickly echoed by two key figures. Musk reposted Amodei's post on the social media platform X and added: "Dario is right." The tech giant, who has repeatedly described AI as a "civilizational-level risk," took a position consistent with his long-standing stance this time, but the timing was intriguing: just a few days earlier, he had dismissed warnings from an Anthropic internal researcher about AI harms as a "conspiracy" and "psychological warfare." OpenAI chief Altman wrote on X: "I agree with Dario. We need to control the pace of frontier AI advancement." Altman's support went beyond words. He said clearly in an interview that OpenAI will not hold an IPO in 2026, citing the severe current AI safety situation and saying that "going public now would be unwise." Altman said on September 14 that the company supports establishing a unified federal AI safety framework to set consistent safety requirements for frontier AI laboratories, and said that "no degree of U.S. competitive pressure can be a reason for acting recklessly." However, U.S. President Trump expressed strong dissatisfaction with Amodei's remarks on Monday, saying that in business history he had never seen an industry leader voluntarily call for regulation, and that once such regulation is strictly enforced, it will make companies disappear or even go bankrupt. He said that a series of negative claims, such as that AI will take over the world and destroy humanity, are a scam. Trump said his administration has already stopped people in the AI industry from engaging in "nefarious or potentially harmful things," and cited Anthropic's Amodei as an example, accusing him of now pretending to be a "perfect little angel." He stressed that the U.S. government already has strong criminal and regulatory powers sufficient to restrain these AI companies and will continue to take action in the future. NVIDIA Corporation CEO Jensen Huang also responded on Monday to the AI safety concerns that have sparked widespread discussion recently. Huang said that warnings from AI company insiders and "whistleblowers" should be taken seriously, and specifically praised Anthropic "whistleblower" Jacob Coxon, saying he had "great courage" in raising concerns and warning that AI companies may be developing too quickly. However, Huang rejected the prediction that "AI may destroy the world," saying such views have "no scientific basis." He said that if AI companies believe their own AI development has already "gotten out of control," then those companies should pause the relevant work or slow the pace of development. In addition, well-known short seller Michael Burry, known as "The Big Short," questioned the AI industry giants' unanimous call to slow the pace of AI technology development, calling the push to slow things down "self-serving." He said that the push by OpenAI, Anthropic, and other large AI companies could slow down fast-growing competitors, thereby benefiting incumbent companies. He also questioned the basic premise of the argument, saying that large language models are not artificial general intelligence, and therefore "there is no AI that needs to be slowed down at all." Burry also pointed out that the so-called "voluntary call to slow down" is merely meant to cover up the real predicament that industry growth has slowed beyond control and IPO plans have been forced to be postponed. In addition, an IPO needs hype and boasting. Shouting "we are too powerful and therefore dangerous" is essentially a form of high-level marketing and self-promotion.