So you’ve got a sketchy email and you want to know if it’s a scam. Or you just got a password-reset link and you’re not sure what to do with it. Makes total sense to ask an AI chatbot, right? Just paste the link in and let it figure things out.
Here’s the problem: anything you paste into a chatbot — links, email content, account codes, all of it — gets sent to that company’s servers to be processed. And according to the Mozilla Foundation, nothing on the internet is 100% secure or private. Your request gets handled in the cloud, and it may be stored by the company. That’s not a glitch. That’s just how this works.
Password-reset links are especially risky. Those links are typically designed to let whoever has them access your account — often without needing your password. If you paste one into a chatbot and that data gets stored, logged, or ever accessed by the wrong person, you’ve essentially handed over the keys.
And it’s not just about hackers. Stanford HAI points out that AI developers have been scraping and collecting massive amounts of data, and researchers have found almost no studies examining how these chatbots actually handle the personal information people share. Hundreds of millions of people are using these tools. The privacy research just hasn’t caught up yet.
The rule is pretty simple: don’t share anything with a chatbot that you wouldn’t want a stranger to read. That includes links from private emails, password resets, login codes, bank notifications, anything like that. IT security experts say this applies even when you’re doing something totally normal — like using AI to help with your resume. You can describe your experience without dumping your full name, address, phone number, and Social Security number into the chat.
If you get a suspicious email, the safer move is to go directly to the company’s official website and check your account there. Or Google the email sender. Or ask someone you trust in person. The chatbot doesn’t need to see the actual email to give you general advice about phishing scams.
AI tools are genuinely useful. This isn’t about avoiding them — it’s about knowing what not to hand them.