Chapter 0.4Chapter 0.4. Critical Thinking and Judgment in the AI Era

In everyday life. You want 12 × 50, but you type 12 × 5 into a calculator. It shows 60, at once and with no doubt. The calculator did its job. Noticing that 60 cannot be right is your job. AI works the same way, on a much larger scale. It writes a full report in seconds, and it sounds sure. Deciding whether that report is right is still your job. It is now the most valuable part of the work.

What you will learn. One habit, the checkpoint. Before you act on anything AI produced, you stop and ask four questions:

1. What exactly is being claimed?
2. What is it based on?
3. What is being assumed?
4. What would change the conclusion?

Then you name who decides, and you write it down.

Six skills make the checkpoint work. You will learn to:

• tell a claim from its evidence
• find the assumptions in a question and in its answer
• notice common reasoning mistakes, which this chapter calls traps
• use AI to find the weaknesses in its own answer
• decide when you cannot be certain
• explain your decision on one page

Why it matters. In a 2026 experiment, people solved logic puzzles with a chatbot that was sometimes wrong. When it was wrong, they got about 31% of the puzzles right. People with no chatbot at all got about 46%. The people with the chatbot also felt more confident (0.4.6). A model sounds the same whether it is right or wrong (Chapter 0.1). So the value is now in the checking. Employers have noticed. PwC, a large accounting and consulting firm, is training more than 85,000 people in this one skill (0.4.1). Chapter 0.3 named Discernment and Diligence. This chapter teaches the thinking underneath them.

The case you will follow. Dave is a controller, the manager in charge of a company's accounting. He asks an AI assistant to confirm a plan: pay every vendor early, to get small discounts. Seconds later he has a confident memo that promises savings of about $42,264 a year. Ana, an analyst at the company, asks for five minutes before he sends it to the CFO, the company's top finance manager. You will follow what she does with them. In the lab, you do her job with the real invoice data.

What you will have at the end. A one-page decision note built from your own decisions, with its numbers checked against the real invoice file. You will also have a habit for the rest of the book.

What this chapter cannot do. It teaches the method and gives you a first practice. Judgment itself grows as you use the method on real work (0.4.1).

Short on time? Read 0.4.1 to 0.4.3 and do Steps 1 to 4 of the lab. The full chapter has seven lessons and a lab of about 1 hour 20 minutes.