Gabriele Mazzini, a researcher at MIT who participated in drafting the European Commission's Artificial Intelligence Act, analyzed why US tech giants frequently create AI panic. The article argues that tech giants are creating "AI out of control" fe

2026-09-22

Gabriele Mazzini, a researcher at MIT who participated in drafting the European Commission's Artificial Intelligence Act, analyzed why US tech giants frequently create AI panic. The article argues that tech giants are creating "AI out of control" fears partly to reshape their image. In recent years, the US tech industry has faced regulatory and public scrutiny due to issues such as privacy violations and adolescent mental health. Now, corporate executives are publicly warning of AI risks, portraying themselves as "guardians" and demonstrating their social responsibility. The article points out that AI systems will not develop will, rebel against their creators, or destroy humanity as depicted in science fiction. The so-called "AI out of control" is more a product risk caused by shortcomings in corporate design, testing, security restrictions, and access management—a risk stemming from human decision-making rather than a threat of machine autonomy. The article argues that in the fierce competition for venture capital and market dominance, establishing "guardrail" mechanisms will slow down deployment. Promoting the "uncontrollability" of technological power is costless and generates billions of dollars in free exposure, thus supporting stock prices. Simultaneously, this approach creates an image of technology as both capable of pioneering cutting-edge technologies and ultimately controlling them. Tech giants' strategy is to push governments to establish higher entry barriers, ensuring that only well-funded large companies can afford compliance costs, thereby consolidating their industry dominance. The article argues that governments should not be swayed by corporate-created panic to drive regulation, but should instead base rules on actual evidence. AI risks certainly need to be addressed, but the focus should be on product liability and safety testing, rather than anthropomorphizing AI as an uncontrollable threat. The real risk is not AI "awakening," but allowing software developers to relinquish human responsibility for the systems they build and facilitating easy market entry. (The above views are summarized from the Wall Street Journal and are for reference only, not investment advice.)