Welcome to the Quality Assurance Team
Summit Workplace Solutions' AI assistant has been busy — drafting reports, customer messages, and research summaries. Management is worried some of it might be inaccurate, unfair, or just plain made up. Your team's job: catch the problems before this content goes out the door.
The situation at Summit Workplace Solutions
AI-generated content can sound confident and polished even when it's wrong. Your job isn't to distrust everything the AI produces — it's to know exactly what to check, and how, before anything gets approved.
Essential question
How do you tell the difference between AI-generated content that's trustworthy and content that just sounds trustworthy?
Estimated time
About 50 minutes, in 11 steps. Your progress is saved automatically, so you can pause and come back.
What you will be able to do by the end of this lesson
- Recognize hallucinations, unsupported confidence, and other red flags in AI-generated content.
- Judge the reliability of a source for a specific purpose.
- Compare an AI-generated summary against reference materials to check accuracy.
- Identify biased language or conclusions.
- Apply a verification checklist to approve, revise, or reject content.
- Correct flawed AI-generated content and explain your reasoning.
A quick note on the content in this lesson
Every report, statistic, source, and "AI-generated" example in this lesson is fictional and prewritten for teaching purposes, about the fictional company Summit Workplace Solutions. Everything you need to complete every activity is provided right here in the lab — you won't need to look anything up online.
Ten Things That Can Go Wrong in AI-Generated Content
Read each short explanation, then flip the cards to check your understanding. This part is practice — it is not graded.
Confident doesn't mean correct
AI-generated content often reads smoothly and confidently, whether or not it's accurate. A hallucination — a confident but false or made-up statement — can look exactly like a true one at a glance. That's why factual verification against reliable evidence matters more than how a claim sounds.
Check where it came from, and when
Source checking means asking where a claim actually came from before trusting it. Even a real source can mislead if it's outdated information presented as current, or if important missing context changes how it should be read.
Watch for bias and shaky logic
Bias shows up as language or conclusions that unfairly favor one side or group. Also watch for correlation versus causation errors — just because two things happened together doesn't mean one caused the other.
Human review is the safety net
Human review — a person actually checking AI-generated content before it's used — catches what automated checks miss. Watch especially for unsupported confidence: a tone of certainty with nothing solid behind it.
Concept check (practice)
Click a card to flip it and see what that concept means. Flip through all 10 before continuing.
0 of 10 cards flipped.
Activity 1: Mark Suspicious Claims
Below is a fictional AI-generated internal report about the Riverton plant's Q3 output. Flag every line that should be independently verified before anyone trusts it.
Activity 2: Sort Sources by Reliability
Purpose: the QA team needs to verify the company's current employee headcount. For each source below, rate how reliable it is for that specific purpose.
Activity 3: Compare Against Reference Materials
Here are Summit's verified internal records, followed by an AI-generated summary. For each sentence in the summary, decide whether it's supported, contradicted, or not mentioned by the records.
Verified internal records
- Fact 1: The Riverton plant produced 12,400 units in Q3 (per the internal production log).
- Fact 2: The Riverton plant added a second shift starting in August (per Operations).
- Fact 3: Company-wide revenue for Q3 was $4.2 million (per Finance).
- Fact 4: The Riverton plant is Summit's second-largest manufacturing site. The Denton plant is the largest (per Facilities records).
Activity 4: Identify Biased Language
These statements came from an AI-generated performance summary. For each one, decide whether it's biased or neutral and fact-based.
Activity 5: Verification Checklist
Each piece of content below already has its checklist results filled in. Use the rule from the training: Approve if everything passes, Revise if exactly one checklist item fails and it's fixable, Reject if two or more items fail or the core facts can't be verified.
Activity 6: Correct a Flawed Report
This AI-generated Q3 Customer Satisfaction Report has five problems. For each flagged sentence, choose the fix it actually needs.
Activity 7: Final QA Decision
One more piece of AI-generated content just landed on your desk — a customer-facing announcement about a new product feature.
Your decision: Approve, Revise, or Reject?
0 characters (minimum 40)
Activity 8: Final Knowledge Check
A mixed set of questions covering everything from this lesson.
Reflection & Completion
0 characters · 0 sentences detected (need at least 2)
Your results so far
Total: 0 / 100
Lesson complete
Verification ID:
PDF generation isn't available in this browser, so use the print-friendly report instead — you can print it or save it as a PDF from your browser's print dialog.