When AI Agrees, Be Careful

I often use AI to think through career decisions.

It's incredibly efficient, and more than once it has helped me see possibilities I hadn't considered before.

But the more I use it, the more I notice something unsettling:

AI can be a little too good at agreeing with you.

How AI Mirrors Your Own Thinking

Ask questions like:

"I feel switching careers might be better than staying where I am. What do you think?"

"Should I start my own business?"

"Is this company the wrong fit for me?"

And surprisingly often, the answer aligns closely with your existing beliefs.

Not because AI is intentionally misleading you.

But because your wording, assumptions, priorities, and even emotions are shaping the conversation long before the answer appears.

What makes this even more dangerous is that AI tends to sound equally confident whether it's right or wrong.

What to Delegate and What to Keep

So the real question isn't whether we should use AI.

It's understanding what decisions we can delegate—and what we shouldn't.

My rule of thumb is simple:

AI is excellent at generating information, structure, and scenarios. It is far less capable of making judgments about your reality or your values.

Here are a few things I happily delegate to AI:

• Exploring options I haven't considered • Breaking down complex decisions into clearer dimensions • Playing devil's advocate • Resume reviews, interview preparation, and other execution-heavy tasks

Where I Stay Cautious

But there are areas where I remain cautious:

• Assessing market realities such as compensation trends, hiring demand, or industry conditions. These require validation from multiple sources, not just AI. • Making major life decisions. AI doesn't know your financial situation, family responsibilities, or what risks you're truly willing to take. • Evaluating who you are. Questions like "Am I suited for leadership?" are often better answered by people who have actually worked with you.

I've come to think of AI as a highly intelligent advisor who has never met me in person.

It can offer frameworks, surface possibilities, and challenge assumptions.

But the final judgment still belongs to me.

The Real Risk: Validation, Not Error

What worries me most, however, is not that AI might give me a wrong answer.

It's that it can make my existing beliefs sound increasingly convincing.

When a response feels perfectly aligned with what I already think, the most important question may not be:

"Is this correct?"

Instead, it might be:

"What would need to be true for the opposite conclusion to make sense?"

The Best Thinking Partners Challenge You

The best thinking partners don't always validate us.

They occasionally make us uncomfortable.

And that's often where the real learning begins.