Critical Thinking: The Human Skill AI Can't Replace

Build critical thinking and decision-making skills to evaluate AI outputs and make decisions that hold up at work.

Role Play

New

Created byStarweaver Experts, Paul Siegel

Last updated 6/2026

English

What you'll learn

Apply critical thinking skills to evaluate AI-generated outputs and judge whether they are reliable enough to act on.
Use analytical thinking to diagnose bias, hallucinations, and reasoning flaws with a structured AI Output Trust Score.
Apply the Bias Stack to spot cognitive, dataset, and model bias before it shapes a real decision.
Combine critical thinking, structured reasoning, and four human-AI workflows to make fast decisions without losing rigor.

Course content

6 sections • 29 lectures • 2h 7m total length

Module Introduction1:44
Introduction to the module, key topics to be covered, and call to action.
The Strategist vs the Technician6:16
Explains the shift from doing the work to evaluating the work, and why evaluation skill now commands a premium in AI-saturated workflows.
System One and System Two at Work6:04
Introduces Kahneman's two-system model and shows where AI use commonly bypasses deliberate thinking in everyday decisions.
The Decision Quality Crisis6:03
Reviews industry survey data on mass AI adoption without review, and frames the financial, operational, and reputational consequences of decision-quality decay.
Judgment in the AI Era0:29

Cognitive Bias and the AI Mirror6:37
Explains how confirmation bias, anchoring, and availability are amplified by AI outputs that reinforce user assumptions.
Dataset and Model Bias6:52
Examines how training data and modeling choices shape AI output before any user interaction occurs, including representation gaps and proxy variables.
Bias Stack in Action: The Amazon Hiring Case6:35
Walks through the documented case in which an AI hiring tool penalized resumes containing markers correlated with female applicants, and traces the bias through each layer of the stack.
Where AI Bias Comes From0:25

The Five Trust Score Criteria6:58
Defines the five criteria: truthfulness, source traceability, relevance, reasoning quality, and risk class, with worked examples from common workplace outputs.
Scoring an AI Summary: The Board Pre-Read6:54
Walks through scoring an AI-generated summary of a board pre-read across all five criteria and identifies where subtle errors would slip through unchecked.
Hallucinations and the Mata v. Avianca Case7:00
Reviews the documented federal-court case in which fabricated citations produced a sanction, surveys common hallucination patterns, and walks through a citation audit checklist any professional can adopt the day after the course.
When AI Invents the Facts0:27
Should this AI-drafted brief go out?
Hands-on-Lab: Diagnostic lab - Bias Stack, Trust Score, and citation audit0:11
The Trust Score and spotting errors

Module Introduction1:46
Introduction to the module, key topics to be covered, and call to action.
Premortem and Red-Team Thinking6:50
Introduces the premortem technique and the red-team approach as complementary methods for surfacing failure modes before a decision is locked in.
Base Rates and Calibration6:40
Reviews reference-class forecasting and Tetlock-style calibration as habits that reduce overconfidence in AI-supported decisions.
Frameworks in a Financial Projection Scenario6:55
Applied walk-through of running a premortem and a base-rate review on an AI-generated revenue projection, and producing a defensible recommendation.
The Premortem0:22

Draft-and-Verify and Adversarial Review7:03
Defines two foundational workflows that combine speed and rigor in AI-assisted work, and the decision criteria for choosing between them.
Premortem-with-AI and Decision Audit6:50
Adapts the premortem method to AI-assisted decisions and introduces the Decision Audit as a retrospective practice that separates decision quality from outcome quality.
Prompt Risk Review and the Moffatt v. Air Canada Case7:00
Walks through the five-question pre-flight check for privacy, confidentiality, and bias amplification, and reviews the published ruling in which a customer-service chatbot misrepresented company policy.
Who Owns What the AI Says0:25

The Decision Journal and the Defensible Memo6:59
Explains the decision journal as a tool for separating decision quality from outcome quality, and introduces a memo structure that surfaces AI inputs, evidence, reasoning, and recommendation for executive review.
Capstone Walk-Through: From Course to Operating Habit6:56
Applied walk-through of the Capstone Decision Memo using the full course toolkit including Trust Score, Bias Stack, premortem, and decision journal.
The Thirty-Day Practice Plan6:34
Walks through the thirty-day post-course practice plan, including weekly journal cadence, one monthly Decision Audit, and how to convert course content into a standing habit.
Keeping a Decision Journal0:25
Hands-on-Lab: Capstone Decision Memo with Prompt Risk Review0:12
Defend your AI-assisted recommendation
Project: Capstone Decision Memo0:04
Decide and defend: the capstone

Requirements

Basic critical thinking and decision-making experience at work, plus hands-on use of a generative AI tool such as a chatbot or document assistant.

Description

Generative AI is transforming the way we work, create, analyze information, and make decisions. While AI can increase productivity and automate routine tasks, it cannot replace one essential human capability: critical thinking. The ability to question assumptions, evaluate evidence, identify bias, and make sound decisions remains one of the most valuable skills in the modern workplace.

In this course, you will learn practical frameworks for evaluating AI-generated outputs and making better decisions with confidence. Rather than accepting AI responses at face value, you will discover how to assess their accuracy, reliability, and potential risks using structured critical thinking techniques.

You will explore how bias, hallucinations, and reasoning flaws can influence AI-assisted decisions and learn proven methods to identify and address them before they create costly mistakes. The course also introduces effective human-AI collaboration workflows that help you combine AI efficiency with human judgment, ensuring stronger and more defensible outcomes.

Through real-world examples, decision-making frameworks, and hands-on exercises, you will strengthen your analytical thinking, problem-solving, and reasoning skills. You will also learn how to conduct Prompt Risk Reviews, create decision memos, and maintain a decision journal that improves your thinking over time.

Whether you are a manager, business professional, student, consultant, or anyone using AI tools in daily work, this course will help you think more clearly, make better decisions, and build a lasting competitive advantage in an AI-driven world.

By the end of the course, you will have practical tools and frameworks to evaluate AI outputs, reduce decision risk, and apply critical thinking with confidence in both professional and personal situations.

Who this course is for:

Mid-career managers and team leads who want to sharpen critical thinking and decision-making skills and lead AI-assisted work with confidence.
Senior individual contributors using generative AI tools daily who need to evaluate AI outputs, catch bias and hallucinations, and decide with rigor.
Compliance, risk, and governance professionals overseeing AI adoption who want to protect decision quality under audit and executive scrutiny.
MBA students and early-career professionals building critical thinking, and AI literacy for AI-era roles.

Critical Thinking: The Human Skill AI Can't Replace

What you'll learn

Explore related topics

Course content

Why human judgment compounds in value5 lectures • 21min

The Bias Stack diagnostic4 lectures • 20min

The Trust Score and spotting errors6 lectures • 22min

Structured reasoning under uncertainty5 lectures • 23min

Collaboration workflows and prompt risk review4 lectures • 21min

Decide and defend: the capstone7 lectures • 21min

Requirements

Description

Who this course is for: