Course 2 · Module 1
Evaluating Models and Outputs
This module is about judging what an AI system gives you. You'll be able to test a model on work that matters to you, check its claims and citations, and spot agreement-seeking and bias in its answers. These skills matter because an AI answer looks the same whether it's right or wrong, so the checking has to come from you.
Module objectives
- Design and run a small test that shows whether a model is fit for a task you care about.
- Apply a verification routine to the claims, sources, and citations in an AI output.
- Identify signs of bias and sycophancy in an output and adjust how you ask and how you read.
Lessons
- Lesson 1 Evaluating AI: Testing a model on your own tasks You'll see why an AI model can be excellent at one task and poor at a similar one, and why published scores won't tell you which is which for your work. You'll be able to build a… 61111h 3m
- Lesson 2 Evaluating AI: Verifying claims, sources, and citations You'll learn how AI systems produce citations and facts that look right and aren't, and how professional fact checkers verify a claim quickly. You'll be able to apply a short… 61111h 3m
- Lesson 3 Evaluating AI: Bias, sycophancy, and persuasion in outputs You'll look at three ways an AI answer can lead you astray while being fluent and polite: by agreeing with you, by carrying a bias, and by being more convincing than its evidence… 611111h 33m