What is AI?

You separated useful definition pieces from sci-fi baggage. “Artificial intelligence” here means: programs that perform jobs people associate with human judgment — recognizing, predicting, translating, summarizing, recommending — often by learning patterns from lots of examples. It does not require a metal body, secret plans, or feelings. Alexa telling you the weather and a factory camera spotting a defective bottle are both in the tent.

A working definition you can say out loud:

AI = software that does judgment-like tasks, often by spotting patterns in data, instead of only following a short list of if-then rules a person typed by hand.

Examples that fit:

• Gmail suggesting a reply ending (“Sounds good, thanks!”)
• A banking app flagging a purchase that looks unlike your usual pattern
• Chat tools drafting an email outline from a messy bullet list

Examples that don’t (by themselves):

• Spreadsheet formulas you wrote yourself (=A1+B1)
• A doorbell that rings because a button closed a circuit
• “Smart” in a marketing sentence with no guessing or learning underneath

Narrower words you’ll meet later: machine learning (learns from examples), generative AI (makes new text/images/code). Those are kinds of AI — not replacements for the broad idea.

• Someone asks “is ChatGPT AI?” — yes; so is face unlock on many phones.
• A headline says “AI did X” — ask which task (classify, generate, recommend) and on what data.
• Setting learning goals: literacy names the family; later lanes go into prompts and builds.

People slide between “AI” (broad) and “a chatbot” (one product style). If your definition only fits movie robots or only ChatGPT, it breaks when Maps reroutes you. Keep it task-based: what judgment-like job is the software doing?

Say your one-sentence definition to a sticky note or voice memo. Then replace “AI” with two real tools: one chat, one silent feature (photos search, spam filter). Does the sentence still work?