Reaction
Gemini 4 Argon scores 87.0% on SpeechMap’s controversial-speech evaluation
SpeechMap says its evaluation tracks which controversial topics AI labs let their models help users write about.
TLDR
SpeechMap says Gemini 4 Argon scored 87.0% on its evaluation of whether AI models help users write about controversial topics. It says Google has been among the better-scoring labs in recent releases. The poster views Google’s comparatively consistent scores as a possible sign of a deliberate approach, while SpeechMap’s site says its tests track when models answer, hedge or refuse sensitive requests.
Combined views
—
2 Sources, first seen ago
