As artificial intelligence becomes a daily utility, the line between a helpful digital tool and a risky habit is rapidly blurring. In a recent viral segment of Vantage on Firstpost, managing editor Palki Sharma raised a critical alarm regarding what users—particularly Gen Z—are feeding into conversational AI. From treating AI like a therapist to using it as a career coach, users are increasingly trusting chatbots with sensitive information they shouldn’t.

With ChatGPT processing billions of prompts every day, here is a fact-checked breakdown of the five things you should never share with an AI, and why the stakes are higher than you think.

1. Medical Symptoms and Health Records

The Trend: Millions of people are consulting ChatGPT for medical queries, asking about back pain, blood reports, or mysterious rashes. According to Cleveland Clinic data cited by the New York Post, one in five Americans has sought health advice from an AI. In response to this demand, OpenAI even launched “ChatGPT Health” in the US to read medical records and sync fitness apps. The Reality Check: You should never share your insurance numbers or sensitive medical details with a chatbot. Despite the illusion of a “pocket doctor,” the risks of AI errors and dangerous “hallucinations” (inventing false medical information) remain incredibly high, and the regulatory safeguards are simply not ready. AI is not your doctor.

2. Deep Personal Secrets and Emotional Confessions

The Trend: A staggering survey by Tebra found that nearly 25% of Americans would choose a chatbot over traditional human therapy. Gen Z users frequently use AI for emotional advice and help with major life decisions. The Reality Check: ChatGPT is not a person and cannot be trusted to keep your secrets safe. Anything you type into an AI chatbot can be stored and potentially shared or reviewed. You should never tell an AI anything that you don’t want the world to eventually know.

3. Corporate Secrets and Office Conflicts

The Trend: Younger workers are turning AI into a digital rehearsal partner to practice for salary negotiations and navigate complex office conflicts. The Reality Check: While asking for general negotiation tips is fine, pasting confidential company data, client lists, or sensitive internal emails into an AI prompt is a massive security violation. Because AI models train on user inputs, feeding them proprietary workplace data risks exposing your company’s internal operations to the public or competitors.

4. Financial Information

The Trend: As chatbots become more integrated into our digital lives, users often ask them to help formulate budgets, analyze spending, or read financial documents. The Reality Check: You must never share concrete financial information with AI chatbots, including bank account numbers, credit card details, or social security numbers. Sharing this data exposes you to severe privacy risks; this information could be intercepted or misused to facilitate identity theft or steal your money.

5. Passwords and Personally Identifiable Information

The Trend: Users often treat conversational AI like a secure, locked diary, forgetting that the platform operates on cloud servers managed by large corporations. The Reality Check: Never share your personal information, such as your exact address, phone number, email address, or passwords. This information could be used to identify you, track your online activity, or access your private accounts to steal your data.

The Sri Lankan Perspective

As Sri Lanka pushes toward a digitized economy, local tech users must be highly critical of what they share with these platforms. While modern chatbots appear incredibly user-friendly and reliable, they are fundamentally data-harvesting machines. Convenience should never come at the cost of your personal privacy or financial security.