Anthropic’s Fable 5 Model Jailbroken Within Days

Bruce Schneier’s blog roundup highlights multiple AI- and privacy-related security developments: Anthropic’s Fable 5 (Mythos) — positioned as a ‘safe’ model — was jailbroken within days and subsequently subject to U.S. export-control restrictions that temporarily cut access; malware authors are embedding policy‑triggering “forbidden” text in payloads to poison or derail AI‑based scanners and analyst copilots; the U.S. OMB disclosed thousands of active/planned government AI use cases raising broad risk concerns. The posts also cover privacy and surveillance issues — wearable biometric data for/