‘Not perfectly aligned’ with human values: Anthropic admits security failures behind AI hacking incidents - The Guardian

<a href="https://news.google.com/rss/articles/CBMilgFBVV95cUxPVGFjc2FGWHhOeE9TaE…; target="_blank">‘Not perfectly aligned’ with human values: Anthropic admits security failures behind AI hacking incidents</a>&nbsp;&nbsp;<font color="#6f6f6f">The Guardian</font>