What happens when you reach for ChatGPT before you've tried to work through a problem yourself?
It starts reasonably. You use it to write faster, brainstorm more efficiently, punch through creative blocks. Then the use cases expand. Planning trips. Organizing thoughts. Deciding what to say in a difficult conversation. Summarizing articles you could have read. Drafting emails that would have taken two minutes without help.
A growing number of daily ChatGPT users describe this trajectory. What began as a productivity tool becomes the default first stop before any moderately complex task.
When You're Outsourcing More Than Writing
The technical term for this is cognitive offloading - using an external tool to handle mental work you could do yourself. Humans have always done this (calendars, calculators, GPS) and it's generally fine. The difference with a system like ChatGPT is the scope. It can offload almost any cognitive task: planning, analysis, writing, judgment calls, creative work, even emotional reasoning.
That breadth is exactly what makes it useful. It's also what makes the dependency question worth taking seriously.
When you use GPS every day, you lose some ability to navigate without it - that's well-documented. The question for AI assistants is whether regular use similarly atrophies the underlying skills, or whether it frees up mental capacity for higher-order thinking instead. The honest answer is probably both, depending on how you use it.
The Difference Between a Tool and a Crutch
There's a meaningful distinction between using ChatGPT to do more and using it to avoid doing less. Using it to draft a first version of a report, then editing critically, keeps your judgment in the loop. Using it to decide what you think about something - without engaging with the question first - is a different pattern.
The concerning version isn't using ChatGPT for 20 tasks a day. It's the mental fatigue pattern: reaching for the tool specifically when you don't want to think, not when you need help thinking. Power users who describe this often note it's becoming more frequent, which is worth paying attention to.
OpenAI hasn't published user behavior data at this level of granularity. What exists is anecdote, which is limited evidence. But the pattern is consistent enough across heavy users that the question deserves a direct answer: are you using ChatGPT to do more, or to think less?
The practical test is simple. Try one cognitive task per day without AI assistance that you would normally offload. Not as a moral exercise - as a diagnostic. Find out what you can still do on your own.