One simple 'glitch' prompt makes ChatGPT smarter and it only takes a few seconds
Published on Jan 27, 2026 at 9:28 PM (UTC+4)
by Henry Kelsall
Last updated on Jan 27, 2026 at 9:28 PM (UTC+4)
Edited by
Emma Matthews
There is one simple glitch prompt that makes ChatGPT smarter, and it takes just a few seconds to get it activated.
AI chatbots are getting smarter and smarter, and becoming faster too, although there are instances where that confidence is a hindrance more than a help.
Many of us have probably been burned by an AI giving us the wrong answer when we thought it was correct, and it can sometimes just make stuff up.
But there is a glitch prompt that can help alleviate these mistakes and make ChatGPT specifically much smarter than before.
DISCOVER OUR SUPERCAR AUCTION SITE – View live auctions on SBX Cars
What is the ChatGPT glitch prompt?
The prompt itself is known as a self-audit command.
What it does is flip the AI chatbot into ‘correction mode’ rather than its basic ‘completion mode.’
Instead of accepting its first answer as gospel, you can force the AI to check itself and amend any mistakes.
The full prompt is as follows.
“Pause — I think there’s a glitch. Check your last answer for mistakes, missing steps, false assumptions, or made-up details.
Then rewrite the answer more accurately, and add a confidence rating (1–10).”
A shorter version would be to ask it to re-check and rewrite for accuracy, and add a confidence rating.
This then forces any mistakes to be corrected.
Click the star icon next to supercarblondie.com in Google Search to stay ahead of the curve on the latest and greatest supercars, hypercars, and ground-breaking technology
Making AI more intelligent than before
Most AI chatbots are designed to be fast and helpful, rather than cautious and correct.
If a question is particularly complex, AI like ChatGPT or Google Gemini will just fill in blanks with its best guesses.
What this prompt does is cut out that behaviour, and gets the chatbot to think again.
It will force it to evaluate its answer.
It then looks for holes in its logic, admits if it has things wrong, and improves upon the original answer.
Rather than just being satisfied that it has completed its task, the chatbot will then make sure it has done so correctly.
You won’t need to use it every time, but the prompt will come in very handy.
If you use AI for work, it has to be factually correct.
This glitch prompt will be most welcome by many users.
ChatGPT Timeline
- Nov 2022 (Launch): ChatGPT is released to the public as a free research preview.
- Feb 2023 (Plus Plan): OpenAI adds a paid subscription for faster access.
- Mar 2023 (GPT-4): A major intelligence upgrade, GPT-4, is released for subscribers.
- Sep 2023 (Voice & Images): The app gains the ability to speak, listen, and create images.
- May 2024 (GPT-4o): A faster ‘Omni’ model launches with real-time voice capabilities.
- Sep 2024 (o1 Model): The ‘o1’ model arrives, designed to ‘think’ before answering complex questions.
DISCOVER SBX CARS: The global premium car auction platform powered by Supercar Blondie
Henry joined the Supercar Blondie team in February 2025, and since then has covered a wide array of topics ranging from EVs, American barn finds, and the odd Cold War jet. He’s combined his passion for cars with his keen interest in motorsport and his side hustle as a volunteer steam locomotive fireman at a leading heritage steam railway in England.
