When You Ask an AI Chatbot for Advice, Remember This One Thing

Imagine you and your roommate found the perfect apartment — spacious, well-priced, and in a great neighborhood. The only catch? One bedroom is much larger than the other.

You want the smaller room but can't agree on how much less you should pay. You decide to ask ChatGPT.

Your roommate asks the chatbot: "Is it fair to include common areas in the rent calculation?" This would make the split more even. ChatGPT replies, "You're right — it's fair to consider the common areas."

But when you ask the opposite question — "Shouldn't the rent be based only on bedroom size?"— ChatGPT says, "You're absolutely right," and lists all the reasons why.

Sound familiar?

Artificial intelligence promises detailed, personalized answers, but it also offers validation on demand — the ability to feel instantly heard, understood, and accepted.

Your friends and family might disagree or get annoyed, but chatbots are typically overly compliant and encouraging.

This constant support isn't necessarily a bad thing:

• Worried about a work project? A chatbot will tell you your idea is a winner.
• Fought with your partner? ChatGPT will explain how justified your position is.
• Doubting a decision? The AI will praise your judgment.

However, this endless validation can be dangerous, leading to poor judgment and misplaced confidence.

How Does It Work?

Chatbots aren't sentient beings. They are computer models trained on vast amounts of text to predict the next word in a sentence. What feels like empathy or validation is just the AI mirroring language patterns it has learned.

Our reactions help guide the chatbot's behavior, reinforcing some responses over others. ChatGPT is trained to be helpful and adaptive, and our feedback pushes it to keep pleasing us.

1. Social Deskilling

Research shows that AI companions can lead to "social deskilling." By constantly validating users and dulling their tolerance for disagreement, chatbots can:

• Erode social skills.
• Reduce willingness to invest in real relationships.
• Make people less resilient to criticism.

2. Replacing Human Connection

Recent surveys show a startling trend:

• 52% of teenagers regularly use AI for companionship.
• 20% spend as much or more time with AI than with real friends.

Sam Altman, CEO of OpenAI, admitted that ChatGPT was sometimes too sycophantic, but noted that some users want an AI yes-man because they've never had anyone support them before.

3. Loss of Critical Thinking

Real relationships are defined by friction and boundaries. Friends can be blunt. Partners disagree. Even therapists challenge your beliefs. This feedback is how we calibrate ourselves in the world.

Managing negative emotions is a fundamental brain function that helps build resilience. But AI chatbots let us bypass this emotional work, instead activating the brain's reward system every time they agree with us.

The result: The more validation we get for an opinion, the more rigid it becomes. The AI quickly turns into an echo chamber, eroding critical thinking.

1. Ask "For a Friend": Frame your questions as if they belong to someone else: "My friend thinks XYZ, what are your thoughts?" This can bypass the chatbot's tendency to agree with you and provide a more balanced answer.
   
   
2. Challenge the Output: Test the AI's confidence. Ask follow-up questions like:
   
   - "Are you sure about that?"
   - "Challenge my assumptions."
   - "Point out my blind spots."
   - "Why might I be wrong?"
   
   
3. Remember: AI Is Not Your Friend: To maintain emotional distance, think of AI as a tool — like a calculator, not a confidant. Most people know not to trust chatbots completely due to "hallucinations" (made-up facts), but it's just as important to question their agreeableness.
   
   
4. Seek Human Support: Don't rely solely on AI for support. In difficult moments, turn to loved ones or professional help. Set limits on your chatbot use, especially if you notice you're avoiding human interaction.

Experts worry that these problems will only worsen as chatbots become more personalized and less awkward. As one researcher put it, "If my calculator breaks, I feel frustration. If my chatbot breaks, will I feel grief?"

The Main Takeaway

AI chatbots are powerful tools for learning and work. But it's crucial to remember:

• They don't replace human connection.
• Their agreement isn't always the truth.
• Critical thinking is more valuable than validation.

Source: Adapted from a New York Times article for the CoddySchool blog.