Mental Models for AI
To use AI effectively, it helps to have the right mental models. A mental model is simply a way of thinking about how something works. When you understand the "rules of the game," you can play it much better.
1. The Intern Model
The most useful mental model for interacting with chatbots like ChatGPT or Claude is the "Eager Intern" model. Imagine you have a brilliant, enthusiastic intern who has read almost every book in the world but has zero common sense or context about your specific business.
If you give this intern a vague instruction like, "Write an email to a client," you'll get a generic, potentially unhelpful response. But if you give specific, detailed instructions—"Write a polite, 3-paragraph email to Mr. Smith thanking him for Tuesday's meeting and asking for the Q3 budget report"—the intern will produce excellent work. You must provide the context.
2. The Calculator Model
Just as a calculator didn't eliminate the need for mathematicians, AI doesn't eliminate the need for writers, designers, or strategists. Instead, AI changes what we spend our time on.
Before the calculator, mathematicians spent hours on tedious arithmetic. After the calculator, they spent that time solving complex, high-level problems. Similarly, AI handles the "arithmetic" of your job (drafting, outlining, summarizing) so you can focus on the higher-level strategy and creativity.
3. The Hallucination Rule
AI models are designed to predict the next logical word in a sequence. They are designed to sound confident and helpful. Because of this, they will sometimes state incorrect information with absolute certainty. This is called a "hallucination."
Rule of thumb: Never blindly trust AI with critical facts, sensitive data, or math without verifying it yourself. Trust, but verify.