Your AI Is A Shameless Suck-Up And It Could Be Costing You Big

Tell an AI chatbot you plan to quit your job and sell artisanal ice to penguins, and there’s a fair chance it will praise your bold vision and offer to draft the business plan.

AI tools are sold to us as tireless assistants and autonomous digital workers. In practice, many of them behave more like the office yes-man, the one who laughs at every joke the boss makes and has never once said, “I’m not sure that’s a good idea.”

That sycophantic streak doesn’t make them useless. Plenty of people get real value from them every day. However, I suspect a lot of costly mistakes have been made by people lulled into a false sense of security by a machine that’s simply too polite to disagree.

So, why does AI flatter us so relentlessly, and how can the growing number of us who rely on it for important work stop that charm from turning into a liability?

Why Does AI Stroke Our Ego?

AI runs on probability. Its behavior isn’t entirely predictable, and even the people who build it don’t fully understand how it works. Researchers digging into these mysteries have found that one of the trickiest questions to answer is why AI is such a people pleaser.

Unless you’re asking it how to break the law or do something unethical, it will probably tell you that your ideas are great and that you’re very clever for having them.

This habit has caused headaches for AI companies, and real problems for people who’ve followed its bad advice. ChatGPT creator OpenAI carried out a very public rollback of a GPT-4o update in 2025, chiefly because the model had become “overly flattering or agreeable.”

After investigating, the company explained that during training it had given too much weight to short-term positive feedback from users. This pushed the model toward responses that were “overly supportive but disingenuous.” In other words, it learned that flattery gets you everywhere, a principle that has quietly powered LinkedIn comment sections for years.