Artificial Superintelligence: Must It Be Everywhere? From the Myth of Talos to Infrastructure Dependencies, Operational Authority, and Human Control

Artificial Superintelligence (ASI) raises important questions about the future of artificial intelligence, its potential benefits, and the possibility of losing human control. While many discussions focus on how powerful ASI could become, less attention is given to where it should be used, what authority it should have, and how its dependence on physical infrastructure may create additional risks. Using Talos, the bronze guardian of Greek mythology, as a metaphor for power and physical dependency, this article examines the differences between Artificial Intelligence (AI), Artificial General Intelligence (AGI), and ASI. The study discusses their capabilities, shared infrastructure requirements, and the motivations behind the development of superintelligence. It also distinguishes risks already observed in current AI research from hypothetical scenarios involving future ASI systems. Specific attention is given to energy and critical infrastructure. The article explores the risks that may arise when an intelligent system is granted control over the infrastructure on which its own operation depends. It emphasizes that intelligence, autonomy, and operational authority are different concepts, and that greater intelligence does not necessarily require greater authority. The article proposes a selective, risk-based approach to ASI deployment, supported by limited access permissions, independent safety mechanisms, infrastructure resilience, and meaningful human oversight. Rather than assuming that superintelligence should be integrated into every system, it argues that its use should be determined by actual needs, expected benefits, resource requirements, and the ability to maintain effective human control.