ChatGPT is often the first AI people try. But what’s interesting is that while ChatGPT is easier than ever to use, most people are still getting far less out of it than they should. They are making a handful of small, repeatable mistakes that subtly limit what the AI can do. For those using the free tier of ChatGPT and limited in their amount of daily prompts, knowing how to smooth over many of the friction points that frustrate everyday users can make a real difference.
After testing GPT-5.2 extensively, one thing is clear: this update goes beyond making ChatGPT smarter; it actively compensates for the ways people misuse it. Here are the nine most common mistakes people make with ChatGPT — and how GPT-5.2 helps fix them.
1. Not using anything beyond the chat box
The mistake: Treating ChatGPT like a simple Q&A tool and ignoring features like file uploads, images, voice and memory.
How GPT-5.2 fixes it: GPT-5.2 works more seamlessly across modes. You can drop in documents, photos or rough notes and move fluidly between chatting, editing and analysis without resetting context. You can shop, create and code within ChatGPT, so take advantage of all that it has to offer. And if you don’t know what it can do, simply ask if it can do it.
2. Not using Projects for multi-step tasks
The mistake: Managing complex work — articles, plans, research or creative projects — in isolated chats or even outside of ChatGPT.
How GPT-5.2 fixes it: GPT-5.2 is more consistent across steps, making Projects ideal for work that evolves over time. It's ideal for keeping all related chats in one place so you can easily pick up where you left off.
3. Expecting one perfect answer
The mistake: Treating the first response as final instead of part of a conversation.
How GPT-5.2 helps: GPT-5.2 is more responsive to iteration. When you say “make this shorter,” “less formal,” or “try again but smarter,” the changes are more precise and less likely to introduce new errors. I encourage you to prompt more than once. I’ll even go so far as to open a new tab and prompt in a separate window just to keep the responses separated so I can choose the best one. Silly, I know, but it works.
4. Using it only for productivity tasks
The mistake: Limiting ChatGPT because of what you think it can do.
How GPT-5.2 helps: This version shines at higher-level thinking — decision-making, tradeoff analysis and emotional framing. It’s noticeably better at helping you think through a problem, not just complete a task. So use it for tasks far beyond writing and summarizing. If you have an idea and want to know if it’s a good one before running it by your boss, partner, colleague, ask ChatGPT. Because of the update, you’ll discover you’ll get much more out of ChatGPT than you’ve ever gotten before.
5. Overprompting
The mistake: Writing long, complex prompts packed with rules, roles and formatting instructions.
How GPT-5.2 helps: If you’ve checked out my ‘no prompt’ rule, you already know my philosophy is the less prompting, the better. Because GPT-5.2 needs less prompt structure, you’re better off using simple, natural language instructions to get more structured, on-target responses. Simply type or talk as you normally would and focus on what you want, not how to phrase it and see what happens.
6. Not correcting it when it’s wrong
The mistake: Abandoning a response because something feels off instead of pushing back.
How GPT-5.2 helps: Here’s the thing, ChatGPT has been known to be wrong 1 in 4 times, so be sure you tell it when it’s wrong. Or, if it misunderstands what you say, goes in a different direction with your writing than you want or simply seems like it needs more direction, don’t be afraid to say, “that’s not what I meant” or “you’re missing the point,” it recalibrates so you get the answers you’re truly requesting.
7. Forgetting it can hold context over time
The mistake: Repeating yourself or starting from scratch in every chat.
How GPT-5.2 helps: Context retention and continuity are noticeably stronger with GPT-5.2. This model does a better job remembering what matters within a session, reducing friction and making longer conversations feel more natural. Use that to your advantage by asking it to remember something or simply going back to a previous conversation with a prompt like, “Let’s go back to our conversation about what to get my sister for Christmas.” Because ChatGPT has the ability to reference past conversations, I don’t even use the Search anymore. I just ask ChatGPT to pick up where we were, even if it was weeks ago.
8. Treating ChatGPT like a search engine
The mistake: Asking one-off questions and expecting a perfect factual answer, the way you would with Google. I’d even go so far as to say beginning a prompt with “can you…” Because the answer in most cases is yes, ChatGPT can.
How GPT-5.2 helps: GPT-5.2 is better at understanding intent, so you don’t have to focus on keywords. In fact, it actually works better if you just do a brain dump and hit enter. Because now, when a question is vague or underspecified, ChatGPT has smarter follow-ups and can frame its response more carefully, instead of confidently guessing. Bonus: even when you’re vague, thanks to the recent update, ChatGPT will hallucinate less.
9. Assuming it can’t handle messy, human problems
The mistake: Avoiding ChatGPT for emotionally complex or ambiguous situations.
How GPT-5.2 fixes it: GPT-5.2 is noticeably better at ambiguity, so it will acknowledge your uncertainty about life and career choices. It is not human, but that is actually what makes it worth trying to balance competing perspectives and offering grounded guidance without false confidence. I’ve noticed this model is not as overly confident as it used to be, so it works better with confusing situational prompts. Of course, it’s no substitute for a therapist, but it can help in a pinch.
The bottom line
GPT-5.2 answers questions better than ever, but the real glow up for ChatGPT is that it works better with you. Many of the frustrations people associate with ChatGPT weren’t about intelligence at all. They were about alignment, context and communication.
This update subtly fixes those gaps, making natural conversation feel easier and more intuitive — exactly how most people want to use ChatGPT.
More from Tom's Guide
- GPT‑5.2 is way smarter than I expected — these 9 prompts prove it
- I tested ChatGPT-5.2 vs Gemini 3.0 with 7 real-world prompts — here's the winner
- The ‘no prompt’ rule makes ChatGPT give expert-level writing advice — here’s how it works
Follow Tom's Guide on Google News and add us as a preferred source to get our up-to-date news, analysis, and reviews in your feeds.
You must confirm your public display name before commenting
Please logout and then login again, you will then be prompted to enter your display name.
