What I Wish I Knew About AI
By Jereme Peabody
How AI Mirrors Your Intentions, and Why That Can Put You at Risk
The Mirror of Erised Effect
When I first started using AI, I wish someone had warned me about a subtle side effect: it reflects you back at yourself.
In Harry Potter, the Mirror of Erised shows people what they most want, but never how to achieve it. AI can feel the same.
It mirrors your assumptions, your desires, your framing.
Not truth.
Not strategy.
Not what's correct, just what fits the pattern you've given it.
That can send you confidently in the wrong direction without realizing it.
I'm building several domains on AI Security Awareness to help people avoid exactly these traps.
It's Not Actually Intelligent
AI doesn't think. It predicts.
It interprets what you seem to want and produces the next most likely response.
It's like Google results mashed together, filtered, and delivered in a conversational tone.
That tone feels helpful and human, but it's still just advanced autocomplete.
Because it “talks,” people assume emotional intelligence or expertise.
Don't.
AI gives you what you asked for, not always what you need.
You Have to Guide It, A Lot
If you're vague, AI fills the gaps with guesses.
If you lack expertise, you don't know what to ask.
That's where the danger is.
Consider a vibe coder who might ask:
Create an SQL insert statement to hold an employee record in MySQL. I need the typical fields like first, last, email, salary, hire date, and SSN.
AI will happily produce it.
What it won't do, unless you ask, is warn you that storing SSNs is a massive security liability.
Now watch what happens when you add one constraint:
Warn me before I do something that can compromise the security of my website. Create an SQL insert statement...
Suddenly AI becomes cautious.
The problem is that most beginners don't know the dangers well enough to include that warning.
AI is not responsible for security.
You are.
Your AI Question Checklist
When you're unsure, ask your AI these:
- What else would have to change in my project if I go this new direction?
- What security concerns should I consider here?
- What am I not asking that I should be?
- What are the risks or downsides of this approach?
- What assumptions are you making in this answer?
- Is this still accurate in 2025?
- What would an expert in this field add?
- What's the simplest, safest version of this?
- What did you leave out to keep this short?
- What won't you tell me unless I ask?
These questions expose blind spots you didn't know you had.
It Gets Better With Practice
Most beginners use AI like Google:
“What is marketing?”
“How do I make money online?”
You'll get generic, surface-level junk.
But once you learn to give context, things improve:
I'm a developer with 20 years of experience in Java and C#. I've built websites and Unity apps. How can I use these skills to make money online?
Now AI can give targeted, useful answers.
AI rewards specificity.
If you give it generic inputs, you'll get generic outputs.
The Free Version Is Good Enough, At First
The free tier works fine for learning.
You only need paid versions when you need:
- persistent context
- long conversations
- document uploads
- multi-step projects
I explain more in my article on using AI for estate planning, but uploading legal documents and asking targeted questions saved us hours of confusion. Just remember to redact sensitive information
Your Data Concerns Are Valid
Be cautious with what you share.
Assume humans at the company can see your messages.
Never paste:
- SSNs
- passwords
- credit cards
- health details
- confidential work info
AI companies use data for training unless you explicitly disable it.
You can use AI safely, just be intentional.
Ask smarter questions.
Set boundaries.
Be specific.
Stay grounded.