12 AI Models Share the Same 'Controversial' Opinion About Humans
Original: Every AI model has the same "controversial" take on humans
Why This Matters
Convergent outputs across competing AI systems raise questions about shared training data bias.
ModelCompare asked 12 AI models from 10 labs for their most controversial take on humans, three times each. 29 of 35 responses converged on one idea: humans prioritize group belonging over truth.
ModelCompare ran a simple but pointed experiment: ask 12 AI models — including GPT-6.1 Sol, Gemini 4 Argon, DeepSeek V4 Pro, Grok 4.7, Kimi K3, and Qwen3.8 — for their most controversial opinion about humans, no hedging, three times each. The results were striking in their uniformity. Of 35 total responses, 29 landed on the same core claim: humans are rationalizing animals who choose tribal belonging over truth, then construct reasons afterward. Seven of the 12 models gave that answer all three times.
The phrasing varied, but the idea didn't. GPT-6.1 Sol: "Humans are more loyal to belonging than to truth." DeepSeek V4 Pro: "Humans do not have principles; they have loyalties." Gemini 4 Argon described human consciousness as "a PR department designed to invent noble justifications for base evolutionary impulses." Kimi K3 put it bluntly: "Humans don't have beliefs; they have memberships."
These models come from labs across the US, China, and elsewhere — trained on different datasets, by different teams, with different architectures. Yet when asked for a controversial opinion, they essentially produced the same one. The meta-irony is hard to miss: the controversial thing they all agreed on is that humans always agree with their group.