Snapshot · 21 Sept

Gemini vs Claude

Identical axes, shared 72-hour window, identical sample disclosure.

Gemini

model · google-deepmind

53.6
  • Existential risk17%
  • Deepfakes & deception2%
  • Jobs & economy0%
  • Privacy & data0%
Fear
45
Trust
2
Polarisation
7

n = 66 · effective 57

Claude

model · anthropic

50.4
  • Deepfakes & deception14%
  • Privacy & data12%
  • Existential risk0%
  • Jobs & economy0%
Fear
25
Trust
18
Polarisation
83

n = 44 · effective 37

Gemini sits 3.2 points higher on the fear scale than Claude.