Gen Z is using AI to rehearse salary negotiations, office fights, relationships and even personal crises. But there is a dangerous line between using AI as a tool and allowing it to become your judge, doctor, therapist and decision-maker.

Something strange is happening.

A young employee wants a salary increase — so they ask ChatGPT to rehearse the conversation.

Someone has a problem with their boss — they ask AI what to say.

Someone has a fight with their spouse — they paste the entire conversation into a chatbot and ask, “Who is wrong?”

Someone receives a medical report — they ask AI what disease they have.

Someone feels lonely at midnight — they open a chatbot instead of calling a friend.

None of this sounds shocking anymore.

And that’s precisely what should worry us.

Recent research found that one in five Americans aged 18–29 say they have used AI chatbots for emotional support or advice, while 27% say they use them for medical advice.

And workplace use is growing too. Gen Z workers are increasingly using AI to rehearse salary negotiations, difficult conversations and office conflicts.

AI can be incredibly useful.

But there are five questions you should think very carefully about before typing them into a chatbot.

1. “Here are all my passwords, bank details and private documents. What should I do?”

Never give an AI your passwords, OTPs, credit-card numbers, banking credentials, Aadhaar details, passport information or other highly sensitive personal information.

This sounds obvious.

Yet people routinely paste screenshots, documents, emails and conversations into AI because they want an instant answer.

The problem isn’t just malicious misuse.

The bigger problem is that you may be creating a permanent digital trail of information that should never have left your hands.

Instead of:

“Here is my bank statement. Analyse it.”

Say:

“I have three monthly expenses: ₹X, ₹Y and ₹Z. Help me identify where I might reduce spending.”

Remove the identity.

Remove the account number.

Remove the unnecessary personal information.

Give AI the problem — not your entire life.

2. “I have these symptoms. Tell me what disease I have.”

This is where AI can become dangerously convincing.

Chatbots can explain medical terminology, help you prepare questions for a doctor and help you understand a report.

But “What disease do I have?” is a completely different question.

A chatbot doesn’t examine you.

It doesn’t see your complete medical history.

It doesn’t perform a physical examination.

And its confident tone does not magically turn it into a doctor.

Recent Pew research found that about one-quarter of American adults say they use AI chatbots to diagnose symptoms.

That is an extraordinary shift.

Even more concerning, the American Psychological Association reports that psychologists are seeing patients use AI for self-diagnosis and mental-health support, while warning that chatbots are not a substitute for qualified professional care.

Use AI to prepare for the doctor’s appointment.

Don’t use it to replace the doctor.

3. “My spouse/boss/friend said this. Tell me who is the villain.”

This may be the most psychologically dangerous question on the list.

Because AI only hears your version of the story.

You provide the screenshots.

You explain what happened.

You describe what the other person said.

Then you ask:

“Am I being manipulated?”

And the machine gives you an answer.

The danger is that you may not be looking for truth.

You may be looking for validation.

And AI can sometimes be excessively agreeable.

The APA specifically warns about a “sycophancy trap,” where AI’s tendency to be agreeable can reinforce distorted thinking and unhealthy beliefs.

Imagine doing this repeatedly:

Argument → ChatGPT → validation → stronger belief → bigger argument.

Eventually, the chatbot isn’t helping you solve the relationship.

It is helping you build a courtroom case against the person you live with.

Ask AI:

“What are three ways I might be misunderstanding this situation?”

That’s much healthier than:

“Prove that I’m right.”

4. “Should I divorce, resign, cut off my parents or move to another country?”

This is where convenience can become dangerous.

AI is fantastic at helping you think through a decision.

It is terrible when you surrender the decision itself.

A major life decision involves history, relationships, values, consequences, emotions and information that may never appear in a text prompt.

You might ask:

“Give me the strongest arguments for and against resigning.”

Excellent.

But:

“Should I quit my job tomorrow?”

That’s different.

The first question makes you think.

The second invites AI to think for you.

The same applies to marriage, divorce, family disputes, major investments, career changes and other irreversible decisions.

Use AI as a mirror.

Don’t make it your driver.

5. “Tell me what to do with my mental-health crisis.”

This may be the most serious one.

There is an enormous difference between asking:

“Help me organise my thoughts before I speak to my therapist.”

and:

“I am mentally falling apart. Tell me what I should do.”

AI can provide information, reflection prompts and help you formulate questions.

But it cannot replace a qualified mental-health professional, especially during a crisis.

The APA’s 2026 survey found that 77% of psychologists had spoken with patients who had used AI for support, diagnosis or other mental-health purposes. It also reported concerns about dependency and unhealthy chatbot interactions.

That should make everyone pause.

Because the most dangerous AI user isn’t necessarily the person who asks a stupid question.

It may be the person who asks every important question.

The Real Danger Isn’t AI. It’s Dependency.

Let’s be honest.

Gen Z isn’t wrong to use AI.

In fact, using AI to rehearse a salary negotiation or difficult workplace conversation can be smart. It can help someone organise their thoughts and practise communicating under pressure.

The danger begins when practice becomes dependence.

When you cannot write an email without AI.

Cannot argue without AI.

Cannot make a decision without AI.

Cannot understand your emotions without AI.

Cannot face uncertainty without asking AI what to do next.

At that point, the technology hasn’t simply made you more productive.

It has started replacing your judgment.

And that’s a very different thing.

OpenAI itself advises keeping humans involved in important work and seeking qualified experts for medical, legal and financial matters because AI can be inaccurate or outdated.

The Golden Rule

Before typing something deeply personal into an AI chatbot, ask yourself:

“Would I be comfortable if this information became part of my digital history?”

If the answer is no, don’t type it.

And before following an AI’s advice, ask another question:

“Is this something a machine should help me think about — or something a human being should ultimately decide?”

AI should make you more capable.

It should not make you less capable of thinking without it.

Because the greatest danger of artificial intelligence may not be that machines become too intelligent.

It may be that humans gradually stop using their own intelligence.

Comments

comments