Friday Squid Blogging: Victims of Unregulated Squid Fishing

Multiple security and privacy risks reported: the US government used export-control authority to classify Anthropic’s Fable model as a dangerous munition and blocked foreign access, causing the vendor to cut access worldwide; malware authors are embedding policy-triggering nuclear/biological text in commented headers to disrupt AI-based scanners and analyst copilots, an active evasion technique that degrades automated threat analysis; the OMB disclosed 3,611 active/planned AI use cases across the US government, raising concerns about large-scale automation of sensitive functions; several mass‑