AI saved me from confidently being wrong

I was writing articles for my business about topics I knew something about. But I kept questioning whether I actually understood them well enough.

Did I know enough to make the claims I was making? Would people care about these issues? Was I seeing the topic the way a reader would?

Answering those questions meant either hiring an expert or spending hours researching each topic. I did not have the budget for an expert, and I did not want to take the time to become one.

I started using AI as a thinking partner.

I would give it what I had written and ask direct questions about my assumptions. Was I thinking about this correctly? Would this make sense to someone with this problem?

One example was an article I started writing about making my own electrolyte powder. I was excited about the idea that citrate could improve mitochondrial energy production because it plays a role in the Krebs cycle.

But after feeding my draft into Claude and asking questions, I realized I did not actually understand the mechanism as well as I thought I did.

Claude explained that citrate is an intermediate compound created inside the Krebs cycle, not something mitochondria directly absorb from food or supplements. It also pointed out that citrate from a drink would be processed elsewhere in the body before ever reaching the mitochondria in that form.

What surprised me was that the original idea was not completely wrong. Citrate still has real benefits, like helping prevent kidney stones and affecting acidity levels in the body. But it was not improving metabolism in the way I assumed it would.

That process completely changed how I approached writing.

Instead of treating AI as a tool to confirm my ideas, I started using it to challenge them.

It saved me from publishing claims that were not actually true.

More importantly, it changed how I learn.

Instead of assuming I understood a topic because I recognized the terminology, I started pressure-testing my assumptions until I actually understood the underlying mechanics.

I was able to write from a more informed and well-rounded perspective, closer to what I would expect if I had consulted an expert.

There is still a tradeoff. I am relying partly on AI’s interpretation rather than learning everything directly, so I fact-check anything that feels surprising.

But overall, I became much more confident in my process and much less likely to mistake familiarity for understanding.

Step 1: Write a draft based on what you already know

Step 2: Share your draft with an AI tool like ChatGPT or Claude

Step 3: Ask it to challenge your assumptions

For example: Am I thinking about this correctly? What might I be misunderstanding? Would this make sense scientifically or practically?

Step 4: Ask follow-up questions until you understand the reasoning clearly

Step 5: Revise your article based on what you learned

Step 6: Fact-check anything that feels surprising or important before publishing