Four things AI does that most people never notice
Guess before you open each one. Every answer links to the full explainer.
?Does a chatbot check its facts before it answers you?
?Will you and a friend get the same answer to the same prompt?
?Can a browser extension read what you type into ChatGPT?
?Does joining a call with an AI note-taker mean you agreed to be recorded?
Three reads that change how you use AI
New to the series? Read these in order. Each one takes 8–10 minutes and ends with checks you can use the same day.
The newest risks, explained
Five questions this series answers
Choose the question that bugs you most. Each track runs from the basics to the deeper cuts.
Why does AI get things wrong?
How models "think", and the blind spots that come with it.
Who decides what you see?
Feeds, rankings and recommendations, and how to take the wheel back.
What does AI know about you?
Chats, meetings, extensions and "free" apps: where your data goes.
- The safe way I use AI tools
- What chatbot data collection reveals
- Why AI answers differ for each user
- Extensions that can read your prompts
- AI note-takers at work: privacy risks
- 7 ways AI tracking follows you online
- Free AI tools: paying with your data?
- The dark side of free AI tools
- 5 dark UI tricks in AI apps
Is it real or AI-made?
Deepfakes, cloned voices, fake reviews and scam tools.
Where is all this heading?
The bigger picture: history, agents, companions and the hype cycle.
I read the vendor's own docs and privacy policies, check settings in the apps themselves, and link the primary sources (FTC, OECD, NIST) inside each post. When something isn't published, I say so instead of guessing. Some posts include affiliate links, which support the site at no extra cost to you (disclosure).
Two free tools I built that run in your browser and store nothing.
Get the next explainer before the algorithm buries it
AI explainers, tips and new Behind the Algorithm posts in your inbox. No spam.