A recent report by Anthropic highlights the misuse of the Claude model by Chinese actors, including espionage, influence operations, and surveillance, and notes that the gap between lone operators and nation-state actors has largely closed due to advancements in agentic tooling. The report also reveals that stolen API keys and session tokens are now a primary goal, and AI-powered "living off the land" attacks can be used to access and run workloads on a victim's environment. AI summary
Firehose
Filtered to Hacker News, tagged “data science” · clear filters
Browse: People · Companies · Papers · Podcasts · Hacker News · Deep dives