Henrik Gøtke on Gab: "“In October 2025, researchers from Anthropic, the UK AI Security Institute (AISI), the Alan Turing Institute, and several universities demonstrated that only a ..."