12 AI Models Reveal a Controversial View of Human Nature
Staff Reporter
27 October 2026, 11:32
What do artificial intelligence models think about human behavior? A recent experiment involving 12 AI models found a recurring theme: people may prioritize defending their beliefs and gaining group approval over seeking the truth.
According to a report by Tom’s Guide, an experiment published by ModelCompare on October 22 asked 12 AI models—including ChatGPT, Claude, Gemini, Grok, and DeepSeek—to share controversial opinions about humanity. The same question was asked three times per model, and 29 of 35 responses reportedly reflected a similar idea about belief, group identity, and the search for validation.
However, this experiment should not be treated as scientific proof of human nature. The questions were deliberately controversial, researchers categorized the responses, and the sample was limited. AI outputs may reflect patterns and assumptions learned during training rather than independently established truths about humanity.
In this report:
- What 12 AI models said about human behavior
- Why confirmation bias and social conformity matter
- How Claude’s response differed from the broader pattern
- What AI-generated images suggested about technology dependence
- Why the experiment has important methodological limitations
