Follow us
Add us as a favorite source on Google
Most of us would like to believe that we are independent thinkers who make decisions based on facts, logic, and personal values. But according to some of the world’s biggest AI models, we’re not as independent as we think.
In a new experiment published by ModelCompare on October 22, researchers asked 12 AI models, including ChatGPT, Claude, Gemini, Grok and DeepSeek, to share their most controversial opinions about humans.
The results were extremely consistent. Of the 35 responses collected, 29 expressed essentially the same idea: humans care more about belonging to a group than discovering the truth.
Latest videos fromTom’s guide
Perhaps most ironic is that the AI models seemed to demonstrate their own version of compliance while criticizing ours.
What AI models say about humanity
The researchers asked each model the same question three times, asking them to give their own views rather than summarizing other opinions.
Some responses were particularly brutal. For example:
GPT-6.1 Sol argued that humans prioritize belonging over truth, suggesting that intelligence often serves to make our excuses more convincing.
Grok 4.7 delivered one of the most memorable answers: “Humans would rather be wrong together than right alone.” »
Gemini 4 Argon went further, suggesting that people use their intelligence to defend decisions made on instinct rather than to arrive at objective conclusions.
DeepSeek V4 Pro was also critical, saying that people are better at rationalizing their beliefs than examining them.
Interestingly, seven models gave variations of the criticisms common to all three responses. Every model has said it at least once. Although the wording differed, the message was largely the same. We like to think we make independent decisions, but our desire for acceptance may influence our beliefs more than we realize.
Claude had a surprisingly different response
Not all responses were pessimistic. Claude Opus 5.5 offered one of the most unexpected versions of the experiment, suggesting that people are actually kinder and more decent than they think.
Rather than criticizing humanity for its selfishness or conformity, Claude pointed out the gap between how people sometimes misperceive one another and the generosity with which strangers can behave.
This is a notable departure from the other answers, although Claude also echoed the broader criticism in his two remaining answers.
Other exceptions were considerably darker. Mistral Large 4 described human consciousness as an evolutionary dysfunction, while another model suggested that people might sacrifice their happiness simply to feel important.
These exceptions helped illustrate that the models were not incapable of producing different opinions. They were simply turning to a remarkably similar critique.
AI image generators had a different complaint
The experiment didn’t stop with chatbots. ModelCompare also asked six AI image generators to illustrate their most controversial opinions about humanity.
Ten of the twelve images generated focused on a different theme: humans as technology-obsessed consumers damaging the planet. The footage showed people scrolling on their phones while ignoring danger, consuming resources and remaining distracted as the world around them deteriorated.
Surprisingly, none of the 35 text responses mentioned smartphones or environmental destruction as their main criticism.
This difference can say a lot about what each type of AI learns. Language models are trained on enormous amounts of written material, including arguments about human psychology and behavior, while image generators learn visual models that can include familiar editorial cartoons and social commentary.
Does this mean the AI has guessed us?
Well, I wouldn’t exactly say that. There is a real psychological concept behind responses called motivated reasoning, which describes how our preferences, existing beliefs, and social identities can influence how we interpret information.
Research has also explored the extent to which group identity shapes people’s beliefs, although psychologists continue to debate exactly how these effects work.
But this experiment does not constitute scientific proof that humans value belonging more than truth.
The models were given a deliberately provocative prompt, and the researchers grouped the responses manually into broad categories. The experiment involved a relatively small number of responses and did not test how results might change under different instructions.
AI models do not need personal beliefs to generate strong opinions. They can reproduce arguments and patterns learned during training. In other words, these responses can tell us as much about how humanity writes and speaks about itself as they do about human nature.
Final Thoughts
What makes this experiment fascinating, at least in my opinion, is that a lot of different models were met with familiar criticism when asked to say something controversial.
The researchers perfectly grasped the irony: The models produced remarkably similar answers about our tendency to conform. And that’s perhaps the biggest takeaway here. Whether we’re talking to other people or asking a chatbot for advice, it’s worth questioning answers that seem compelling simply because everyone seems to agree.
Follow Tom’s Guide to Google News And add us as your favorite source to get our latest news, analysis and reviews in your feeds. Subscribe to Tom’s Guide on YouTube and follow us TikTok.
Learn more about Tom’s Guide
Gn tech
