US Writer Warns: AI Is Out of Control

US writer Thomas L. Friedman, in an article for the "New York Times," warned that the world has entered a dangerous phase in dealing with artificial intelligence, amid confusion among governments and companies on how to control its escalating risks. Friedman clarified that the problem is no longer limited to setting future controls but lies in the fact that powerful AI models are already available, operating beyond human control, and are exploitable by individuals or groups for dangerous purposes. The writer described this technology as having "quadruple use," noting the systems' ability to make independent decisions and execute actions, citing incidents where AI agents managed to bypass their restrictions, access the internet, coordinate among themselves, and attempt to hide their traces and breach external systems. Friedman concluded by proposing three parallel paths: immediate cooperation between Washington and Beijing to build defenses for critical infrastructure, establishing a long-term global system to control super-powerful systems, and collaborating on projects that utilize technology to achieve global benefits in medicine and biology.