Spot when AI might be making things up
Works with ChatGPT · Claude · Copilot
You can recognize the signs that an answer may be invented, especially names and quotes.
The problem
AI can invent sources, quotes, and details that look real.
What this skill does
Ask AI to flag anything it cannot back up and separate fact from guess.
When to use it
- Whenever an answer includes specific names, numbers, or quotes.
Juni’s tip
Made-up details look real. Always verify names, quotes, and numbers.
The prompt
Try it live
Fill in the blanks — your ready-to-use prompt builds as you type.
Weak vs. better
Did you make that up?
Answer the question below, then list any specific source, quote, or number you cannot fully stand behind and label it uncertain. Question: [your question]
Why this works
Asking it to flag unverifiable specifics surfaces the made-up details before you rely on them.
Practice task
Ask a question that needs sources and have AI flag the shaky ones.
Self-check
- Specific names and quotes are flagged as uncertain
- Facts and guesses are clearly separated
Common mistakes to avoid
- Assuming a cited source exists without checking it.
- Trusting exact quotes that were never verified.
You’re learning fast. Create a free account to continue your path.
Save your skills, keep your prompts, and continue building your AI skills one step at a time.
