'Straight up lies': Anthropic CEO attacks OpenAI's US military deal (www.techradar.com)

Anthropic CEO Dario Amodei has issued a strong rebuke of OpenAI's recent announcement regarding its new partnership with the Pentagon, accusing the company of "straight up lies" in its messaging about the deal. This controversy follows Anthropic's decision to withdraw from a government contract over safety concerns surrounding AI applications in domestic surveillance and autonomous weapons. In contrast, OpenAI's deal, which it claims incorporates more "guardrails," has faced backlash from users, leading to an increase in uninstalls of ChatGPT and a rise in the popularity of Anthropic's AI bot, Claude. Amodei's internal memo highlights skepticism regarding OpenAI's reassurances, labeling them as "safety theater" that does not adequately address ethical concerns. He questions the vague stipulation of "all lawful use" within the Pentagon contract, pointing out the potential implications for surveillance practices. As Anthropic continues discussions with the government, the dynamics between AI companies and military partnerships remain contentious, raising critical questions about the industry’s ethical responsibilities and the reliability of AI safety measures.