When Your Audience Isn't Who You Thought It Was (and What AI Taught Me About It)
TL;DR
I assumed my library presentation audience had basic marketing knowledge—but that assumption was wrong. The feedback (mostly from one person) forced me to rethink how I explain AI. Ironically, asking ChatGPT for a "cave-level" explanation produced the clearest, most human-friendly definition of AI I've seen yet—and it reshaped how I'll approach my next talk.
I'm still thinking about my last presentation at the library.
I got feedback afterward. The librarian told me most of it was positive, but there was also some specific criticism. The feedback itself was a bit scattered, but the core issue was clear: the content didn't fit the audience.
I had assumed I'd be speaking to small business owners with at least a rudimentary understanding of marketing. That assumption felt reasonable—the event description suggested exactly that. But it turned out to be wrong.
The librarian, to her credit, still wants me to give more talks. The next one is titled "What Is AI?" I agreed immediately and told her I'd slow things down and present at a much more basic level.
Then she followed up with an important detail:
All of the negative comments came from one person.
That's always tricky. One voice can be loud, thoughtful, insightful—or simply misaligned with the rest of the room. But it was still enough to make me pause and rethink how I approach audience level, especially in a public library setting where the range of experience can be enormous.
Turning to AI for Help (Again)
As usual, I turned to ChatGPT for help outlining the "What Is AI?" talk.
It gave me solid content—but once again, it was too high-level for true beginners. For example, it suggested explaining AI like this:
Common types of AI you see today
- Generative AI – creates content (text, images, audio, video)
- Predictive AI – forecasts outcomes (sales, churn, risk)
- Recommendation systems – suggest what to watch, buy, or read
- Automation AI – handles repetitive or decision-based tasks
If you already work in tech or marketing, that makes sense.
If you don't? It's meaningless.
So I tried a different approach.
The "Living in a Cave" Prompt (Highly Recommended)
I asked ChatGPT this:
"Assume the audience has been living in a cave with no outside influence for the last 20 years. Explain what AI is."
Here's what it gave me:
What is AI?
AI is a kind of computer program that can learn from examples and make decisions without being told every step.
Simple. Clear. Human.
But this was my favorite part:
A cave-level explanation (which honestly made me laugh):
Imagine teaching someone to recognize a dog.
Old computers:
- You had to explain every rule
- Four legs
- Fur
- Tail
- Ears
- If the dog looked different, the computer got confused
AI:
- You show it thousands of pictures of dogs
- Over time, it learns what dogs usually look like
- When it sees a new animal, it makes its best guess
That guess isn't random—it's based on patterns it learned.
And then it hit me.
This is exactly why people say AI feels "human."
Not because it thinks like us—but because it learns from experience instead of rules, just like we do.
The Real Lesson (for Me)
The biggest takeaway from all of this wasn't about AI at all.
It was about audience calibration.
When you speak to a general audience:
- You can't assume shared vocabulary
- You can't assume context
- You can't assume curiosity looks the same for everyone
And sometimes, the best explanations don't come from dumbing things down—but from changing the frame entirely.
For my next library talk, that's exactly what I'll do.
FAQs
What is the simplest way to explain AI to beginners?
AI is software that learns from examples instead of following rigid, hand-written rules. You teach it by showing it data, not instructions.
Why do people find AI explanations confusing?
Most explanations assume prior knowledge of technology, data, or business concepts. Without shared context, terms like "predictive" or "generative" don't land.
Is AI actually thinking like a human?
No. AI doesn't think or understand. It identifies patterns in data and makes probability-based guesses that feel human because they resemble learning.
How should presenters adjust for mixed-level audiences?
Start with everyday analogies, avoid jargon, and assume zero prior knowledge. You can always layer complexity later.
Is one piece of negative feedback worth changing a talk?
It depends—but even a single comment can reveal a mismatch between expectations and delivery. The key is deciding whether to adjust content, framing, or pace—not abandoning your expertise.