Co-Intelligence — Ethan Mollick

Summary in 15 minutes. We have built something that talks like a person but doesn’t think like one. First, we need to understand it. Mollick’s four simple rules for using AI without fearing it — and without trusting it blindly. Conversational partner, creative collaborator, coworker, tutor and coach — and the futures that may lie ahead. You’ve read the book. Now invite AI in, stay in the loop, and map your own frontier.

Key ideas of "Co-Intelligence: Living and Working with AI"

An Alien Mind

We have built something that talks like a person but doesn’t think like one. First, we need to understand it.

1. Three Sleepless Nights: The moment you realize something has changed
2. Aligning the Alien: Alignment, bias and guardrails
3. The Jagged Frontier: Where AI excels, and where it stumbles

Four Rules for Co-Intelligence

Mollick’s four simple rules for using AI without fearing it — and without trusting it blindly.

1. Always Invite AI to the Table: Rule 1
2. Be the Human in the Loop: Rule 2
3. Treat It Like a Person — but Which Person?: Rules 3 and 4

The Many Faces of AI

Conversational partner, creative collaborator, coworker, tutor and coach — and the futures that may lie ahead.

1. AI as a Person: A machine that converses like one of us
2. AI as a Creative: An idea machine
3. AI as a Coworker: Centaurs, cyborgs and secret cyborgs
4. AI as Tutor and Coach: Rethinking learning and expertise
5. Four Possible Futures: Where AI might go from here

Put It to Work

You’ve read the book. Now invite AI in, stay in the loop, and map your own frontier.

1. The 7-Day Plan: One task a day. Tick it off when done.
2. One-Page Cheat Sheet: The whole book at a glance — save it
3. Final Word — Co-Intelligence: The human beside the machine

From the free first chapter

When ChatGPT arrived in November 2022, Wharton professor Ethan Mollick started working with it in his classroom almost immediately, and quickly saw that it was unlike any software before it. To really understand AI, he argues, you need to lose about three sleepless nights — the point at which it sinks in that this thing can do what you assumed only people could.

Today’s large language models (LLMs) learned from enormous amounts of text to predict which word is likely to come next. Out of that simple mechanism came abilities nobody fully planned — writing poetry, coding, giving advice — that even their builders can’t completely explain.

Read the first chapter free · Price: ৳250 · or every book with a membership