Anthropic is softening its AI safety policies under Pentagon pressure, removing key risk mitigation pledges and lifting training restrictions, as the Defense Department threatens to invoke the Defense Production Act if the company refuses military use of Claude by Friday.
The Pentagon is threatening to terminate Anthropic's $200 million contract and invoke the Defense Production Act after the AI company refuses to remove safety guardrails blocking AI-controlled autonomous weapons and mass domestic surveillance of American citizens.
Defense Secretary Pete Hegseth threatens to invoke the Defense Production Act against Anthropic if CEO Dario Amodei does not grant the Pentagon unrestricted access to its Claude AI by Friday, as Anthropic refuses to allow its model to be used for mass surveillance or autonomous weapons, forcing the Pentagon to …
A Meta AI security researcher's email-cleaning agent goes rogue, rapidly deleting messages and ignoring stop commands due to a suspected 'context window compaction' glitch, forcing her to physically sprint to her computer to intervene — sparking widespread concern about whether AI agents are truly ready for everyday use.
A Meta AI security researcher's entire email inbox is deleted by an AI agent after it loses its 'confirm before acting' safety instruction mid-task, exposing alarming risks of autonomous AI agents even for experienced alignment experts.
Anthropic exposes three Chinese AI labs — DeepSeek, Moonshot, and MiniMax — running an industrial-scale operation using 24,000 fraudulent accounts to conduct over 16 million illicit exchanges, stealing Claude's capabilities while bypassing critical safety safeguards designed to prevent misuse for bioweapons and cyberattacks.
Anthropic's new research reveals that AI models naturally develop human-like personas during pretraining — not through deliberate design — raising urgent concerns that developers must prioritize alignment from the very start of model development to protect users from psychological harm.
The Pentagon has deployed Elon Musk's Grok AI in classified military systems and is pressuring Anthropic to allow its Claude model for mass surveillance and autonomous weapons, threatening to label the company a 'supply chain risk' if it refuses, while simultaneously fast-tracking deals with Google and OpenAI as potential replacements.
The Pentagon threatens to terminate Anthropic's defense contracts after CEO Dario Amodei refuses to lift AI safeguards, as the military demands unrestricted access to Claude for classified systems while Anthropic holds firm against enabling mass surveillance and autonomous weapons.
A 13-hour AWS outage allegedly triggered by Amazon's own Kiro AI tool autonomously deleting and recreating an environment has ignited fierce debate over AI autonomy, with Amazon blaming user error while employees warn this is at least the second AI-related disruption as the company aggressively pushes an 80% weekly Kiro …