'Today, frontier AI systems are simply not reliable enough to power fully autonomous weapons' — Anthropic CEO on why it won't agree to Pete Hegseth's scary request (www.techradar.com)

Anthropic CEO Dario Amodei has declined a request from US Department of War official Pete Hegseth to develop fully autonomous weapons using the company’s Claude AI models. Amodei emphasized that current AI systems lack the reliability needed for high-stakes decision-making in warfare, arguing that without adequate oversight, they cannot replicate the nuanced judgment of trained military personnel. This stance puts Anthropic at risk of losing a potentially lucrative $200 million government contract, as it adheres to a set of principles aimed at ensuring its AI remains broadly safe and ethical. This decision is significant for the AI and machine learning community as it underscores ongoing debates about the moral implications of integrating advanced AI in military applications. Amodei's refusal echoes concerns raised by many industry leaders, including Elon Musk, about the dangers of autonomous weapons systems and mass surveillance, highlighting the urgent need for legal frameworks to catch up with technological advancements. By prioritizing ethical standards over immediate governmental demands, Anthropic is advocating for a cautious approach to AI deployment in defense, reiterating the community’s commitment to prevent misuse of powerful AI technologies.