My GPT with Over 5000 Chats Removed from GPT Store Without Explanation: 1.5 Months of Development and Promotion Wasted

[klym]

UPDATE:
My ‘Video Summarizer AI’ has been reinstated in the GPT Store. Many thanks to the OpenAI team for the quick resolution of its issue, and to everyone who supported me and my GPT during such a difficult moment!

---

---

For the past one and a half months, I have dedicated myself to developing a GPT called “Video Summarizer AI” [ChatGPT - Video Summarizer AI]. The aim of this GPT was to enhance the learning experience by generating educational summaries from user-submitted videos.

This GPT, with over 5000+ chats, received positive feedback from users, some of whom were highly enthusiastic. It secured the 10th position in the Education category. During this period, I personally handled backend development, testing, user interactions, and user experience improvements. Additionally, I invested my funds in marketing and advertising for this GPT.

Yesterday, I was notified of its removal, without any clear explanation. The message vaguely suggested a potential violation of the Terms of Service or Usage Policies. Consequently, I meticulously reviewed these documents and amended any aspect that could potentially be in breach.

Today, I received a second notice of removal, again without detailed reasoning, merely referencing the same documents I had already examined in depth. Even a lawyer friend of mine found no grounds for violation.

When I received the second notice, I was outside and could only glimpse the message title on my phone, as the message wouldn’t fully load. Later, at home, on my laptop, I discovered there was no detailed reason for the removal. Furthermore, I learned that reinstatement in the GPT Store was not an option, nor was there a way to appeal.

My goal was to be a proactive developer, dedicating resources to development and advancing GPTs on the OpenAI GPT Store. This ambition is now unattainable if my projects are removed without specific justification.

Dear Openai, if the decision is irreversible and not a technical error but a conscious choice, I implore you to enlighten me about any mistakes I may have made. I have earnestly sought clarity, even enlisting legal counsel for a thorough review, yet the reason for this decision remains elusive.

[N2U]

In the email you received it states that:

In the near future you’ll be able to submit an appeal

So you might have to wait a bit, but it looks like you’ll be able to appeal this decision

We can’t really handle issues about account, billing and takedowns from the chatGPT team here on the forum, it’s better to reach out though the link in the email, or help.openai.com

[klym]

@N2U, thank you so much for your support and guidance!

I completely understand the process and appreciate the opportunity to appeal in the future. However, I’m concerned about the immediate impact of this removal on my users. They can no longer access my GPT, and I worry about the effects of even a short downtime. If a popular product becomes unavailable, even briefly, it can significantly erode user trust and perception of its reliability. I’m feeling a bit overwhelmed, as I’ve been working tirelessly to enhance the product’s user experience. The news of the removal, especially without a clear reason, came as a shock after all the effort and sleepless nights dedicated to improving the product. Thank you again for your understanding and assistance during this challenging time.

[N2U]

I fully understand where you’re coming from, stuff like this can be frustrating and downright scary, but take a deep breath and rest assured that these things will be solved eventually

I’d definitely recommend that you try recreating your GPT as an “assistant” over at the playground, this will give you an option to put it on your homepage.

I don’t know why your particular GPT has been taken off the store, but I can take a guess if you’d like?

[klym]

@N2U Thank you for your understanding and reassurance. It’s comforting to know that others recognize the challenges and uncertainties in situations like these. Your suggestion to recreate my GPT as an ‘assistant’ in the playground is an interesting idea, and I appreciate your recommendation to feature it on my homepage. That could indeed be a valuable alternative.

As for the removal of my GPT from the store, I am curious about your thoughts and would welcome your guess. Different perspectives might shed more light on the situation, helping me navigate this better. Again, thank you for your support and for taking the time to offer helpful advice.

[cass]

I’m really sorry for your experience. Just a thought. If you take out any mention of YouTube from the description and title and republish would that help?

I think the key maybe they are not allowing titles or descriptions with copyrighted terms?

[N2U]

This is just my personal thoughts on this, but I think some of this is due to concern’s about the possibility of user input being copyrighted.

Legally speaking, I’m allowed to view a YouTube video. That is the entire purpose of the video. The original creator usually retains the actual copyright on the video, while YouTube has non-exclusive rights to publish that video. If I want to publish that video on my own website, I’ll have to contact the original creator or other relevant copyright holders to obtain the necessary permissions.

If someone ask’s a GPT to summarize a YouTube video that they don’t hold the rights to, what happens then?

Is this a violation of copyright? If so, who is liable, the user? the creator of the gpt? OpenAI?

These questions aren’t something that I, OpenAI or anyone else who isn’t a copyright lawyer or judge can definitively answer. But I’m guessing your GPT is on hold while OpenAI figures out how to handle this.

[PromptBreeders]

I experienced the same thing with one of my GPTs :

Honestly, I can’t understand which Policy I could have violated with this GPT, which just has a bit of recent knowledge on the trends of 2024, freely accessible on the internet.
In short, content that OpenAI itself uses to train its models.

If I compare it with some that I see in the GPTstore (like the dozens that give investment advice even though it’s forbidden), it’s incomprehensible…
I suspect reports from the ‘competition’ (automated?).
In any case, this new store is not completely transparent for developers, and the selection of retained GPTs raises questions…

My advice given the traction you’ve had with this GPT: take this censorship opportunity to make it a product in its own right, avoiding the use of the OpenAI assistant (which, from my understanding of the Policies, is likely to suffer the same fate).

[grandell1234]

Adding on to @N2Us response, OpenAI is currently being sued for violating copyright law. I think they may have removed your GPT because it could be used for stealing copyrighted work. I would guess that after the situation with the New York Times has calmed down, they might be more lenient. Another theory is that your GPT is automatically removed if it is flagged enough. So a group of people could go through trending or certain categories and falsely flag a bunch of GPTs to try to get them removed.

[PromptBreeders]

The oldest among us, former plugin developers, probably remember with emotion the ‘A/A#A1A’ war this summer to be listed on the first page of the plugin store

Different times, different battles, I have the impression that the battle of GPT flags/reports is starting…

[klym]

This removal occurred even though the store still hosts copycats and similar GPTs that clearly infringe on the ‘YouTube’ trademark in their names and lack educational purposes, not covered by the ‘Fair Use’/'Fair Dealing" acts, in contrast to my GPT, which remains educational and compliant.

@grandell1234 , you proposed this hypothesis:

Another theory is that your GPT is automatically removed if it is flagged enough. So a group of people could go through trending or certain categories and falsely flag a bunch of GPTs to try to get them removed.

Perhaps this is the real reason, originating from an unfair automated moderation system for GPTs. It is not only unjust but also seems unintentionally designed to ban the best GPTs. Competitors could exploit this, lodging complaints against leaders in various niches, potentially leading to the consistent and unfair banning of good GPTs. If this hypothesis holds true, I would like to urge OpenAI to take immediate action to prevent these unfair, competition-motivated bans.

[klym]

@PromptBreeders about it:

Different times, different battles, I have the impression that the battle of GPT flags/reports is starting…

It’s sad news that the GPT Store system has this vulnerability, and instead of directing people to create additional value, it provides options to destroy others’ creations.

[N2U]

I think it’s best to wait and see what happens before we jump to any conclusions.

the hypotheses of @grandell1234 and I, could both be true, or they could be completely wrong, we don’t actually know yet, although I’ll still recommend submitting an appeal when that becomes possible

[PromptBreeders]

I completely agree with you @klym.
It’s a lot of wasted energy and disappointment for the real creators.
As for me, I will not take part in this battle.

[PromptBreeders]

I agree with you @N2U , these are only hypotheses at this stage.

But these GPT bans, even if they were only meant to be temporary, benefit others (the GPT copycats) in a definitive way: the user base, long and dearly acquired, instantly migrates to these GPT clones.
It is not normal for these bans to be carried out in a brutal manner without prior explanation or recourse.

[N2U]

Who knows, this is all theoretical, but while this can benefit others, it can also hurt them as well, nothing prevents people who are abusing the system from getting perma banned at a later stage, and they’re likely to be subject to the same kinds of abuse if their gpt becomes popular as well

[JD_2020]

Throw my name in the hat. Was a Featured GPT in fact a day ago it was being used so much.

WebGPT: https://x.com/jd_2020/status/1745144260188352630?s=46&t=y4pl0S1-1KmOPIcD7Hq-hA

Greg Brockman even shouted it out:

I realize this is only likely to trigger resentment or jealously. I don’t post it as a flex or to brag. But simply as a testament to how confounding this is. I’ve been a Plugin dev since May of 2023, launched the GPT Store as a top GPT, and somehow got banned.

[JD_2020]

For a fact, mass-reporting of plugins is happening.

[chieffy99]

If I spoke too harshly, I apologize.

I wonder.

How do you become a developer?
The announcement states that if there is a misunderstanding, you will be able to appeal. And the store has only been open for a few days. It is not enough to handle all bugs or problems. If you follow the news, you may find that many people are like you. My GPT disappeared for almost a week and then came back. That’s it be.

Part of your trouble may be a system glitch. But the other part comes from you. If you think carefully and make appropriate preparations. You can’t be concerned about impacting your users when you should have a website to announce the news. Additionally, when developing AI, you should always have a backup of your data. If you do it yourself, you should be able to create a GPT backup to prevent problems like this.

Also, have you sent your mail to OpenAI yet? They’re busy these days but I guarantee you’ll get a reply from a human faster than Go-gl- and he’ll listen to you.

If your GPT is not lost but the next day, it doesn’t respond to the prompts you use every day. When you fix it, the next day it changes again. And again, every day is different. I bet your users will have fun using it. Just GPTs gone, you still can’t handle it. I assureม you don’t want that to happen.

And what if users use your GPTs to violate the rules? Have we ever thought about what it would be like? Did you know that before GPT became aware of the violation? It must receive information into itself every time. I think it’s not just you who is in trouble.

[fabrizio.salmi]

I came from decade YouTube partnership now stopped just because I’m tired of a buggy platform able to let freebooters stream a copyrighted soccer match, publishing NSFW ads and ignoring reports from legit creators.

In addition they blocked me when I uploaded my own music, pure raw techno I made myself with absolutely no copyrighted stuff, they just triggered by API driven upload process and… who knows which others hidden flags.

To be clear I never received a strike in my life with more than 10 channels out there.

For the sake of the story I driven 50M views to collab channels and touched 25M on one of mine then I am not a newcomer in such context and fortunately I have a job then I never relied on YouTube for living.

I also faced a prosecution in my country just because I made a top ranked website including embeddable YouTube content and of course at the end I won but… K€ lost to defend myself, 5 years of long legal battle and if I go now on the platform I can spot thousands of unlegit stuff in less than 10 queries.

I now use Jellyfin with a bunch of cronjobs and it works perfect for my media needs and I support my fav creators outside the platform

The worst on that platform is that you can’t report copyright infringement if you are not the owner.

To be honest I’m automatically documenting thousands per day of unlegit stuff to a private GitHub repo, opened to interested stolen owners, since the platform is not interested into reduce traffic… nor to clean some automated chain of booters and spammers.

Can I say that just 5000 items counts for 4B views? Yes it is.

Some if not most of them were ads powered, loaded by freebooters, clearly unlegit and nothing changed across years and multiple adpocalypses.

I also did streaming on YouTube for top NGOs and I can say a fair moderation there is not an option unless you push strong with a proper, educated legal team.

In my little custom GPT experience when I read that to have GPT in the name is discouraged I immediately changed its name and verified that old links still work, that was the expected behavior to my opinion and I was happy to see that by OpenAI.

Hopefully the appeal process in the GPT ecosystem will be released soon giving to those who seen their GPTs removed a new life.