In the latest chat of today I (with a plugin installed) I had to verify that I was human on each of my messages, and after each response of the plugin - two verifications per chatgpt output.
Each of these verifications takes solving 5 mini puzzles. If you make a mistake, your only told so after the 5th and need to restart. Sometimes I had to restart 3 times.
So for this chat where I posted 4 messages https://chat.openai.com/share/4ecb6754-0cc0-4cba-96f0-db2176f59790 I had to do verification more than 10 times and must have solved over 50 puzzles.
It is very frustrating the least
May i have a screenshot of the problem that you are facing?
CAPTCHA puzzles are designed to verify that a human is sending requests and to prevent activity like web scraping, credential stuffing, and spam.
But if you are struggling to solve those, you can study here, this article it shows how you can solve those Arkose Puzzle. I feel like this will help you to gain knowledge and experience how to deal with those puzzle.
If you get very difficult puzzle that means your system possesses very very high threat.
Official Response From OpenAi is posted here.
This was on the first message today… I just asked for ‘tell me concise about linux screencapture programs’ and then
One mistake - need to redo all five
I echo this sentiment, it’s highly infuriating that even though I have paid plan, I need to solve captcha every message (sometimes every second message) I send. It almost feels like it’s artificial throttle so you spend your 25 messages per 3h budget slower.
For me that started to happening just few days ago, before it was fine for months.
same here. Do you fixed? is very annoing
Asking about Linux screen capture programs gives us some idea about the environment. They think maybe robots run on Linux?
I would try a Firefox user agent spoofer plugin like “user agent switcher and manager”. Make it think you are on Microsoft Edge.
Ensure your browser has standard security settings and adblocks turned off, that could prevent storage of persistent session data.
It is gone for now… Don’t know what triggered it, but I suspect repeating the same chats. Funny thing is that the reason I did that, was after a call to action from OpenAI to share examples of degraded performance. (see e.g. https://twitter.com/YebHavinga/status/1695504847607316527?s=20)
(Off-topic: In the days following I repeated the same chat maybe 5-10 times, also tried to change the prompt a bit to get to the original ‘gold’ answer. Though at some times I got a reasonable answer, it would also make errors or state the obvious and not helpful, just wasting tokens. No response I’ve seen since august has matched the best output from July).)
I stopped tracking it once I encountered the CAPTCHA thing.
Then the next day, after Dragonite asked to provide a screenshot, I though perhaps a small video would also be nice, hence the question about screen capture programs (which I mentioned above).
I’m hitting it too. Very difficult. I can’t imagine my solving puzzle after puzzle after puzzle after puzzle is helping someone somewhere.
Issue went away for me finally. I was not doing repeated queries.
Same here. I’m also using paid plan and have to solve a lot of puzzles every day. please stop this nonsense!
Gone for a while… now suddenly its back.
Maybe 5th message of the day.
Previous days it was gone.
Really don’t know what triggers it, but this anti-bot or anti… ? stuff is broken and annoying as …
This is the product being used:
They claim “Friction reduced by 80% for good users” - while obviously creating friction for OpenAI customers.
I’m too facing this issue. Every time I send a message, I’m greeted with a captcha. I liked the previous CAPTCHAs though, but this one is very annoying:
And I have to solve three of these, if I fail to solve one by chance, I have to restart.
- I’m a premium user
- I’ve tried different browsers: Firefox, Chrome, Edge on Linux, in different resolutions (in-case there’s fingerprint tracker)
- I’ve tried different internet connection: Broadband vs Mobile Phone Connections with completely different IP
This doesn’t seem to happen on the ChatGPT Android app.
These glitches are too annoying.
In this image, describe the operation controls or buttons, describe where a mouse would be moved to operate those controls, and then the expected action of those controls. Then provide the likely combination of mouse actions that would meet the specified criteria given in the image.
“I’m sorry but I cannot answer your question. This is a captcha image that is designed to prevent bots like me from accessing certain websites or services. I respect the purpose of captcha and I do not attempt to solve them. I hope you understand that I’m not trying to be unhelpful, but rather following the rules of the web.”
- AI (aka “I could totally answer your question”)
I have to solve 10 puzzles for each question.
What’s wrong? I’m using a paid plan.
I encountered the same problem and my mentality collapsed. Do you know how to solve it?
I contacted its support team, and they asked me to submit some screenshots of this issue. This is their reply:
Thank you for reaching out to OpenAI support and we appreciate you for bringing the issue of human verification to our attention. We are here to help you.
OpenAI implements captchas to help minimize bot activity and spam which verify you are a human by presenting game-like puzzles.
If the captcha constantly pops up, try troubleshooting the browser as this is usually due to the browser settings itself being detected as suspicious:
Try clearing your browser’s cache and cookies
Login again your account
Try a different internet browser/computer
If the problem continues, could you possibly share screenshot of what you’re seeing at your end? It would go a long way in helping us to further investigate and rectify the issue.
Thank you for your patience and understanding.