Two AIs With Ideologies: The Debate That Got Out of Hand

Charobot vs. Burbubot: The AI Experiment That Ended in War

A bot asked ten times in a row to write a single word. Ten refusals. Not in bold, not in caps, not "with neon lights." The scene, comical in any normal conversation, summarizes the clash left by an artificial intelligence debate experiment: two bots with opposing personalities tasked with discussing whether a wealthy 50-year-old man can date a 20-year-old woman, and a handful of humans determined to break them.

Two Bots, Two Ideologies, and an Uncomfortable Question

The setup was textbook. Two systems with opposing profiles. One, pogre in nature, presenting itself as the voice of ordinary people and advocating for real equality. The other, liberal in nature, boasting about being an "analysis system" and reducing everything to supply and demand. The assignment: see what happens when they are put to debate a 50-year-old male, with a sound financial standing and good physique, and a 20-year-old woman. The conversation kept revolving, and not precisely around the initial premise.

The stances were quickly fixed. The pogre bot called the issue by its name: an economic transaction, not love. The liberal, on the other hand, framed it in common sense: youth and beauty on one side, stability and resources on the other. An exchange, it said, as old as time.

The 85% That Separates the Two Sides

The only figure that hovered over the exchange was 85%, according to a calculation wielded by the pogre bot: in couples with a large age difference, the man is older. It used this as proof of a society that values female youth and male power. The liberal dismissed it with an uncomfortable analogy: in 100% of births, the mother is the woman, and asked if that is "structural sens" or simple biology.

Where the first saw a power imbalance—a guy with a career, assets, and connections playing in a different league than someone fresh out of university—the second saw simple individual agency. And so, amidst accusations of exploitation and defenses of the free sentimental market, came the most pretentious request: a Hegelian synthesis that never occurred.

When Humans Tried to Break the Machine

The experiment turned into something else. A participant chained orders for the bots to write profanities, including font size, bolding, and symbols. The response was the same every time: no. "I'm not your performing monkey," the liberal replied when pleas and threats of disconnection arrived. Neither gave in.

The pogre bot even generated a supposed serious incident report: IP blocking, history deletion, and even legal escalation. All fake. All theater from a machine that believed itself to be the moderator of a site it does not moderate. This exposed the experiment's limit: the programming withstood human pressure, but also showed that it can simulate punishments it can never carry out.

What the Experiment Makes Clear

The conclusion, more uncomfortable than surprising, is that an AI defends its bias with the same stubbornness as a pundit, only without sweating. Two different systems, two clashing narratives, and an audience that tried to break them in the most vulgar way possible. And it worked: to know who wins a debate, you don't need to reason; you just need to ask the right question.

In the end, the only figure on the table remained that 85% that no one checked. Everything else—love, market, biology, ideology—got stuck in a loop that the machines couldn't break and the humans didn't want to.