Five Things You Should Never Tell an AI Chatbot – For Your Own Financial Safety

Intro

AI chatbots like ChatGPT, Google Gemini, and Microsoft Copilot have become everyday tools for answering questions, drafting emails, and even brainstorming ideas. But as these tools integrate deeper into our lives, a Washington Post column this week warned users about a less discussed risk: the sensitive personal data you might inadvertently hand over. According to the column and supporting research, sharing certain types of information with a chatbot can put your money and identity at risk. Here’s what to keep out of the chat window.

What happened

The Washington Post piece, published April 25, 2026, listed five categories of data users should never tell an AI chatbot. It followed a wave of reports about AI vulnerabilities. In February 2026, the BBC reported that a journalist managed to hack both ChatGPT and Google’s AI in just 20 minutes, demonstrating how easily chatbots can be tricked into revealing stored information. Meanwhile, the National Council on Aging (NCOA) in March 2026 highlighted financial scams that increasingly use AI-generated content to target older adults. The common thread: chatbots often retain conversation logs, may use them for training, and can be exploited by attackers if the provider suffers a breach or if the chatbot is manipulated.

Why it matters

When you share financial details with a chatbot, you are essentially trusting that company’s security practices and privacy policies. But those policies vary widely, and data stored in cloud servers is not immune to breaches or legal requests. Even if a company promises end-to-end encryption (many do not), the conversation may still be used to improve the model unless you opt out. If a hacker gains access to your chat history, they could obtain:

  • Full bank account numbers and credit card digits
  • Your Social Security number or driver’s license
  • Answers to common security questions (mother’s maiden name, first pet, etc.)
  • Tax documents or investment screenshots
  • Upcoming travel plans and real-time location

Any of these can enable identity theft, unauthorized transactions, or targeted social engineering. A scammer who knows you’re about to fly to Florida, for example, can send a fake “flight cancellation” text that looks plausible because they know your itinerary.

What readers can do

Follow these five rules to keep your finances safe when using AI chatbots:

  1. Never type full account numbers or passwords. Not even if the chatbot asks for “verification” or offers to help you with budgeting. A legitimate AI tool will never need your login credentials. Use temporary or masked data for testing (e.g., “card ending in 1234”).
  2. Avoid sharing government ID numbers, including Social Security, passport, or driver’s license. These are the keys to your identity. If a chatbot helps you with a form, leave those fields blank or use placeholder text.
  3. Do not disclose security questions or PINs. Even if you think you’re just practicing, that information can be stored and later retrieved. Instead, write them down in a password manager.
  4. Refrain from uploading sensitive financial documents. Screenshots of bank statements, tax returns, or investment portfolios contain far more detail than you realize – account numbers, addresses, employer info. Crop or redact before sharing, or better, don’t share at all.
  5. Keep your location and travel plans vague. You can ask for restaurant recommendations without saying “I’ll be in Chicago next week.” Use general queries like “What are good restaurants in Chicago?” and avoid linking to your specific dates or home address.

Additionally, review your chatbot’s privacy settings. Most platforms let you delete conversation history or opt out of training. Enable those options. Consider using a separate, pseudonymous account for sensitive queries. And remember: anything you type could be seen by a human reviewer or leaked – treat the chat window like a public bulletin board.

Sources

  • Washington Post column “Don’t tell your AI chatbot these 5 things to keep your money safe” (April 25, 2026)
  • BBC article “I hacked ChatGPT and Google’s AI - and it only took 20 minutes” (February 18, 2026)
  • National Council on Aging “Top 5 Financial Scams Targeting Older Adults” (March 17, 2026)