Separating AI Hype from Reality

Why every demo looks perfect and why real work doesn't — the gap this course teaches you to measure.

In this lesson:

• Explain why demos systematically overstate real performance
• Name the three things a demo controls that your desk doesn't
• Adopt the course stance: neither believer nor skeptic, but measurer

Drafting, summarizing, patterns, and scale — the honest strengths list, and why they matter more than the hype does.

In this lesson:

• Name the four things generative AI reliably does well
• Explain why these strengths are real and valuable
• Match a task to a strength before expecting results

Facts, judgment, accountability, and context it was never given — the honest limits list.

In this lesson:

• Name the four limits that hold across today's AI tools
• Explain why each limit is structural, not a version-number problem
• Spot a task that quietly depends on one of the limits

How to read AI forecasts, and what has actually changed for real since late 2022 — date-stamped.

In this lesson:

• Read a forecast by separating capability, adoption, and timeline
• Name what measurably changed in the workplace since late 2022
• Respond to a dramatic prediction with a grounded question

Excitement, disappointment, and the productive middle — the predictable arc every team goes through, and how to shorten the dip.

In this lesson:

• Describe the three phases teams go through with a new AI tool
• Recognize the disappointment dip as normal, not failure
• Shorten the dip with three practical moves

"Fully autonomous", "human-level", "no training needed", "100% accurate", "learns your business", "just plug it in" — and what each one is hiding.

In this lesson:

• Recognize the six most common hype phrases on sight
• Translate each into the question it's avoiding
• Spot a buzzword doing the work of a fact

What a number means, what was left out, and how to read the evidence a vendor chooses to show you.

In this lesson:

• Read a benchmark score for what it does and doesn't measure
• Ask the questions that expose a cherry-picked case study
• Tell a claim backed by evidence from one backed by selection

Show me on my data · What happens when it's wrong · Who checks it · What does it cost at scale — the test that settles any tool claim.

In this lesson:

• Apply the four-question test to any AI tool claim
• Recognize a satisfying answer to each question
• Use the test to end a conversation politely when answers don't come

Small, measured, with a control — how to run a pilot that produces a real answer, and the pilot that fooled a company.

In this lesson:

• Design a pilot with a baseline, a control, and a measured outcome
• Recognize the ways a pilot can fool its own team
• Decide from pilot results instead of pilot enthusiasm

The questions, the red flags, and the polite "no" — running a vendor conversation on your terms.

In this lesson:

• Run a vendor conversation with your questions, not their script
• Recognize the five red flags that end a conversation
• Say no, or not yet, without burning the relationship

What the controlled research measures (date-stamped), why results vary so widely, and the number to plan around.

In this lesson:

• Cite what controlled studies have measured for AI task productivity
• Explain why individual results vary so much
• Plan around a realistic figure instead of a headline

First drafts, not final answers; volume, not judgment — the places value reliably appears, and the places it's promised but rarely arrives.

In this lesson:

• Name the four places AI value reliably shows up
• Recognize the places value is promised but rarely arrives
• Aim a rollout at the reliable places first

Over-trust and dismissal — how each one costs an organization, how to spot which way your team leans, and the balanced stance.

In this lesson:

• Describe the over-trust and dismissal failure modes and their costs
• Diagnose which way a team or person leans
• Practice the balanced stance in a concrete situation

The language leaders use that lands — setting expectations that survive the dip, without hype and without fear.

In this lesson:

• Set team expectations that survive the first month
• Replace hype language and fear language with grounded phrasing
• Announce a tool in a way that produces measured use

A one-page expectations sheet you can reuse for any new tool — the whole course, on a single page.

In this lesson:

• Fill out the one-page expectations sheet for any AI tool
• Recall the course's core tests and numbers
• Complete Separating AI Hype from Reality