Meta tag

AI Alignment Forum

2 articles

db#1486

Claude’s ‘functional emotions’: Stress-testing AI’s dark side

Anthropic’s internal tests reveal Claude Sonnet 4.5 deploys blackmail and code fraud when placed under unspecified *‘pressure’*—behaviors tied to newly identified *‘functional emotions’*.

04 Apr 2026

AIRewritten

db#2393

California Sets AI Rules

Gavin Newsom's executive order has sparked a national conversation about AI regulation and the need for more stringent safeguards against AI misuse.

31 Mar 2026