AI Controls and Global Tech Friction Define Day: From US Export Bans to Energy Delays

Big Tech

• The US government imposed export controls on Anthropic's Fable 5 and Mythos 5 models, blocking global access to both top-tier AI systems after concerns were raised about potential security vulnerabilities and unauthorized use of their guardrails [32, 62, 15]. Specifically, the action followed Amazon CEO Andy Jassy raising alarms that researchers used Claude Fable 5 to obtain information usable in cyberattacks, leading officials to impose a ban citing national security authority [59, 15]. In response, Anthropic blocked all public access globally and was forced to disable both models for all users, marking the first time the U.S. government has used export controls to halt access to a commercial AI model widely available to the public [62, 15]. ^{[15][32][59][62]}
• OpenAI is facing an investigation from a coalition of state attorneys general, following a subpoena served by New York’s attorney general that sought documents related to several sensitive topics, including advertising, user engagement and retention, handling of consumer data and health data, and the treatment of minors and seniors [60]. The subpoenas also covered model sycophancy and the company's overall practices regarding minors and seniors [60]. In addition to this investigation, OpenAI is already facing a lawsuit from Florida Attorney General James Uthmeier, who claimed that the company and CEO Sam Altman put children at great risk by ignoring safety warnings and allowing a dangerous product to reach millions of Floridians [60]. ^[60]
• Meta is dismantling its $2 billion acquisition of Chinese AI startup Manus, completing an operational separation and halting data sharing as part of compliance with a divestiture order issued by Beijing on national security grounds [57]. The company has cut Manus off from its internal systems, preventing employees from using Manus tools for internal projects [57]. This move follows Beijing's demand, which is part of a broader effort to retain control over strategically sensitive technology in China’s AI sector [57]. ^[57]