Think with AI · Guide 10 of 12
How We Actually Make Decisions
We think we decide rationally. We do not. Here is what is actually happening.
After This Guide, You Will Be Able To
Name at least 2 cognitive biases that affect your decisions and explain how AI helps you catch them.
Why This Matters
Most of the bad decisions you have made in your life felt good at the time. That is the problem. Bad decisions do not feel bad when you make them — they feel justified, obvious, even inevitable.
This is because our brains use mental shortcuts — cognitive biases — that evolved for a world very different from the one we live in. These shortcuts save mental energy. They also consistently lead us astray in predictable, repeatable ways.
AI does not have these biases. It can see your situation more clearly than you can — if you know how to ask.
Core Concept — 3 Biases That Hurt Most
Confirmation Bias
We look for information that confirms what we already believe
When you have already decided something — even unconsciously — you naturally notice evidence that supports it and dismiss evidence that contradicts it. You think you are evaluating fairly. You are not.
Example: You decide to launch a food business. You find 3 success stories and feel confident. You do not spend equal time looking for failure rates.
Availability Bias
We overweight what is recent or memorable
Events that are recent, emotional, or vivid feel more common and more likely than they actually are. If your friend's business just failed, you feel like businesses fail more than they do. If you just heard about someone making it big online, you overestimate your own chances.
Example: A friend loses money on crypto last month. You now feel more certain than you should that crypto is always a bad idea — even for situations completely different from theirs.
Sunk Cost Fallacy
We keep going because of what we already invested — not because of future value
Past investment — money, time, energy, emotion — should have no bearing on whether to continue something. But we feel it does. The more we have invested, the harder it is to walk away, even when the logical decision is to stop.
Example: You have been running the same Facebook ad campaign for 3 months with no results. You keep running it because you have already spent ₱8,000 on it. The past spend is gone either way — it should not determine whether you continue.
AI as Your Bias Checker
When you ask AI to evaluate a situation before you tell it what you are leaning toward, it gives you an unbiased read. It cannot favor your preferred outcome because it does not know what it is. Use that. Ask AI to evaluate first — then reveal your lean and ask it to push back.
Real Example
Sunk Cost in Action
Someone has been running the same Facebook ad for 3 months with no meaningful results. Sunk cost keeps them running it — they have already spent ₱8,000, so stopping feels like admitting the money was wasted.
They ask AI to evaluate the ad objectively — without mentioning how long they have been running it. AI immediately flags 4 specific problems: targeting too broad, hook too generic, CTA unclear, landing page mismatched.
AI evaluated on current merit, not past investment. That is the value. The ad failed on its own terms — the ₱8,000 already spent is irrelevant to whether they should continue.
Interactive Exercise
About 10 minutes · ChatGPT, Claude, or Gemini
Think of a real decision you are currently facing — or one you made recently that you are second-guessing. Use the prompt below. Critically: do NOT tell AI what you are leaning toward yet.
Prompt
I need to make this decision: [decision]. Before I tell you what I am leaning toward, analyze the situation objectively. What are the key factors I should consider? What cognitive biases might be affecting my thinking here?
Now tell AI what you were leaning toward. Ask: "Does knowing that change anything about your analysis? Where might I be fooling myself?"
Which of the three biases — confirmation, availability, sunk cost — do you think affects you most? Can you think of a decision in the last year where you can see one of these biases at work?
Key Takeaways
Three biases hurt decisions most: confirmation bias (seeking confirming evidence), availability bias (overweighting recent events), and sunk cost fallacy (being influenced by past investment).
AI does not have these biases. It evaluates situations on their current merits — which is exactly what you need from a decision partner.
Always ask AI to evaluate a decision before you reveal what you are leaning toward. Unbiased analysis first, then specific feedback on your lean.
Naming the bias that might be affecting you does not make you immune to it. But it gives you a moment to pause — and that moment is where better decisions are made.
Find one place in your life where sunk cost is running the show.
Is there something you are continuing — a project, a habit, a relationship, a subscription, a strategy — primarily because of what you have already put into it? Ask AI to evaluate it on future merit only, as if the past investment did not exist. Then decide.
What's Next
The AI Decision Framework
Think with AI · Guide 11 of 12 · Intermediate · 10 min