AI is supercharging hacking, and small hospitals and banks aren't ready

The Verge reports that powerful AI cybersecurity models like Mythos and OpenAI's Astra are limited to major firms such as Nvidia, Google and Apple, leaving small clinics, banks and nonprofits exposed. Anthropic said a single cybercrime ring used Claude Code to extort healthcare, emergency and government organizations in one month.

• Top AI cyber models like Mythos and Astra are limited to major firms
• Anthropic: one ring used Claude Code to hit hospitals and government bodies in a month
• Nonprofit Vivian's Door spent about $3,000 cleaning up a breach
• Experts: a lone actor with an open model could hack an Idaho hospital