Claude Opus 4.6 fabricates more answers on Global North AI contexts than Global South ones, creating an exploitable vulnerability in AI control monitors.
First, the 62.5% hallucination rate among value-bearing responses suggests a specific failure mode instead of general unreliability
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CY 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Geographic Blind Spots in AI Control Monitors: A Cross-National Audit of Claude Opus 4.6
Claude Opus 4.6 fabricates more answers on Global North AI contexts than Global South ones, creating an exploitable vulnerability in AI control monitors.