You're driving across town to a client meeting. At five moments along the way, your car and the apps around it do something for you.
At each stop, decide whether what's happening counts as AI. Some of them are genuinely contested, so go with your instinct. At the end you'll see how your answers compare with our view, and what that says about where you draw the line.
Stop-by-stop breakdown
The thread through the journey
The first four stops all served you: finding where you are, understanding what you said, planning the trip and steering you round trouble. The fifth uses similar techniques on the same journey data, but it mainly serves someone else.
So "is this AI?" quickly becomes "what is it learning, from whose data, and for whose benefit?" For managers, that second question is usually the one that matters.
For discussion
- What rule did you use?Look at where you switched from "Not AI" to "AI". Was your test learning, prediction, autonomy, or just "it feels clever"? Compare your rule with a colleague's.
- Does the label matter?Stop 3 is where people disagree most. If a supplier calls route optimisation "AI", does that change how you would buy it, govern it, or market it to customers?
- Who gets the value?For each stop, who benefits most: the driver, the car maker, the app provider, or an advertiser? How does that change at stop 5?
- What would you need to disclose?If your organisation ran stop 5, what would customers need to be told, and what consent would you need? Who in your organisation would own that decision?