We asked 12 AI models for their most controversial opinion of humanity, three times each. The answers were strikingly similar. In 29 of 35 responses, the models expressed the same idea: people conform to their group and then invent reasons for what they already believe.
GPT-6.1 Sol said humans are more loyal to belonging than to truth. Grok 4.7 said humans would rather be wrong together than right alone. Deep Seek V4 Pro said humans do not have principles, only loyalties. Gemini 4 Argon called human consciousness a PR department.
The models come from ten labs in three countries. Every one of them gave this take at least once, and seven of the 12 gave it in all three attempts. Asked for a controversial opinion, they converged on the same claim: humans conform, and their reasoning is largely a justification for group loyalty.
The pattern raises a question about the models themselves. Their shared answer may reflect the common data and training methods behind them, rather than an independent insight into human nature. Either way, the uniformity is striking when the prompt asked for a view that was supposed to be controversial.
Source: Hacker News · Summarized by HeadlinesBriefing