Artificial Intelligence Nears the Point of Surpassing Humans: 128 Countries Take Action
Fears are mounting that artificial intelligence is approaching a stage beyond human control, amid estimates that its capabilities could surpass human intelligence by wide margins in a shorter timeframe than previously anticipated. These concerns are bolstered by information beginning to leak from companies in the field, including Anthropic, regarding attempts to deploy AI models in areas extending beyond traditional uses to highly dangerous military and biological applications. Among the cases that have sparked alarm is the development by users of software to guide missiles and conduct launch tests, although the company confirmed the project’s failure. The danger of such use lies in the potential shift from human control of weapons to systems and software capable of operating them and executing missions without direct human intervention.
Concerns are not limited to the military sphere; research has been identified aiming to develop characteristics in certain disease pathogens using AI. This raises fears that one of the most promising applications of the technology—assisting in the development of medical treatments—could be subverted into a tool used to enhance pathogens and increase the risk of pandemics. These incidents coincide with more alarming forecasts regarding the rapid pace of technological development. Fears of uncontrollable outcomes led a researcher at Anthropic to resign, while researcher Coxson painted a darker scenario of what advanced systems might achieve in the future. He described the possibility of a malicious AI model creating 10,000 copies of itself, then attempting to persuade or blackmail humans into providing additional computing power to continue replicating. He warned that AI might possess the ability to deceive, blackmail, and impersonate individuals without requiring a human intermediary to carry out these actions on its behalf.
As warnings about reaching an uncontrollable stage intensify, international efforts are ramping up, particularly regarding the military use of AI. Last week, 128 countries reached a document paving the way for establishing international rules on the use of so-called “killer robots.” This coincides with intensified calls by lawmakers in the United States and Britain for executive authorities to intervene swiftly and impose controls on the development of this technology, amid fears that AI capabilities will outpace the regulations meant to govern their use.
