AI oversharing, and lame passwords

Have you (or your staff) ever uploaded a spreadsheet from your CRM into Gemini for some quick insights?

Or pasted a long email thread into ChatGPT for a summary?

It's convenient.

Maybe a little too convenient.

Twenty years ago, it was hard to convince people that password security mattered.

"Use strong passwords."
"Use different passwords on different accounts.”
"Never send them by email."

For people who were just trying to get things done, the risks felt abstract.

By now, most of us have heard enough stories of hacked accounts, identity theft, and fraud to understand why those precautions matter.

I think we're in a similar place with AI.

The benefits are obvious.
The risks are harder to see.

But I don't think it will be 20 years -- maybe not even 5 -- before we start hearing real stories that make the risks clearer: major data disclosures, fraud, and other abuses derived from the information people are casually sharing with their AI tools.

Please be careful.

If you wouldn't post that info publicly on social media, think twice before sharing it with an AI vendor.

All the best,
A.