US AI arms race: Can America stay ahead of China as Pentagon struggles to control military AI risks?
AI is becoming a national security weapon as the US faces China in a new AI arms race, raising fears over cyberwarfare, weapons and AI safety.
The United States is embroiled in a new AI arms race, but the government struggles to maintain control. The military, intelligence agencies, and defense contractors are rapidly integrating artificial intelligence into various aspects of national defense. This development became evident when the US Air Force ordered defense contractors to remove Anthropic's AI products, but the Pentagon later reversed the directive.
The dispute highlights the difficulty Washington faces in regulating AI while still leveraging its most potent tools.
The situation intensified when Anthropic CEO Dario Amodei expressed concerns over restrictions on the use of its technology by the military. Amodei demanded assurances that the products would not be used for domestic surveillance or to develop fully autonomous weapons without human control. Despite these conditions, the Pentagon banned Anthropic's products from use in weapons and control systems.
This ban inadvertently created a predicament when Anthropic unveiled Mythos, a powerful AI system capable of discovering weaknesses in computer networks, making it crucial for cyberwarfare. The National Security Agency (NSA) continued testing Mythos, arguing that its absence would hinder the identification and resolution of vulnerabilities in US military systems.
The NSA's experiments with Mythos for both defensive and offensive cyber operations pose a strategic challenge for the United States. Banning such a powerful AI tool could compromise America's cyberwarfare capabilities against rivals like China, Russia, North Korea, and Iran. The technology could significantly alter the cyberwarfare landscape, potentially giving China an advantage due to its rapid advancements in AI.
This dilemma underscores the ongoing debate between safety and speed in AI development. Supporters argue that stricter rules could cost America its six-month lead over China, while critics fear that moving too quickly could lead to the proliferation of dangerous systems that bypass human oversight.
The Trump administration's recent discussions with AI company executives, including those from OpenAI and Anthropic, aimed to review new models for potential risks involving biological weapons, classified US systems, and communications networks linked to nuclear weapons. However, the policy exempts open AI models from this review process due to their customizable nature.
This exemption is concerning because many Chinese AI companies also develop open models, potentially putting the United States at a disadvantage. The controversy surrounding Anthropic exemplifies the broader issue of dependence on the very companies the government seeks to control.
Written by urgent.news from Hindustan Times - World News's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
