Forum Copilot - Automatic Spam and NSFW Image Detection with OpenAI ChatGPT

Forum Copilot - Automatic Spam and NSFW Image Detection with OpenAI ChatGPT 1.0.1

No permission to download
Hi @Yenxji - what do you mean by the blurry images? Do you mean instead of moving the content to Approval queue, it instead just blurry the images?

For the third party API, I guess you mean you want to use other non-OpenAI API? And the reason being the OpenAI is too expensive or blocked in your region?
I am not talking about spam; I would like you to create an add-on that could detect NSFW images and blur the image of attachments or screenshots, profile pictures, and resource icons only. Because everyone can check their own profile settings > Content Blocking,

Choose what type of content you want to block on the forum site.

Adult content:
Yes / No

Show adult content: Yes / No

Blur adult attachment images: Yes / No

Blur adult resource images: Yes / No

Blur adult media images: Yes / No


Yes it's too expensive, if this is not possible then how about include category and resource list to NSFW add-on and user group view permission?
 
The API does not invoked from the user's IP, so unless your server is placed in a region that is blocked by OpenAI, I don't see why you need a proxy for this?
Maybe I don’t understand something, but isn’t the request coming from the server’s IP? the user has nothing to do with this, the request comes from the server, depending on what country it is in. That’s why I asked, if you don’t understand, then let’s skip the question, I was just asking.
 
I am not talking about spam; I would like you to create an add-on that could detect NSFW images and blur the image of attachments or screenshots, profile pictures, and resource icons only. Because everyone can check their own profile settings > Content Blocking,

Choose what type of content you want to block on the forum site.

Adult content:
Yes / No

Show adult content: Yes / No

Blur adult attachment images: Yes / No

Blur adult resource images: Yes / No

Blur adult media images: Yes / No


Yes it's too expensive, if this is not possible then how about include category and resource list to NSFW add-on and user group view permission?
So essentially you want to let your members to choose if they want to view a full non-moderated images or not, by having a settings in the user profile? If user okay to see NSFW images, just show to the users, and vice versa?
 
Maybe I don’t understand something, but isn’t the request coming from the server’s IP? the user has nothing to do with this, the request comes from the server, depending on what country it is in. That’s why I asked, if you don’t understand, then let’s skip the question, I was just asking.
I would be interest to know which region of server running xenForo forum software are currently blocked by OpenAI? I am not aware of any unless you are placing the servers in China etc?
 
I would be interest to know which region of server running xenForo forum software are currently blocked by OpenAI? I am not aware of any unless you are placing the servers in China etc?
I think it doesn’t matter where the server is located, Vietnam, China, Singapore, Russia, I asked a simple question is it possible to specify a proxy, you never answered, which means that as I understand it, there are no settings. Why inflate the holivar, the question was simple.
I have several plugins where by default there are proxy settings, if they are needed, the author was just worried about whether they are needed or not, you don’t, further discussion makes no sense.
 
I think it doesn’t matter where the server is located, Vietnam, China, Singapore, Russia, I asked a simple question is it possible to specify a proxy, you never answered, which means that as I understand it, there are no settings. Why inflate the holivar, the question was simple.
I have several plugins where by default there are proxy settings, if they are needed, the author was just worried about whether they are needed or not, you don’t, further discussion makes no sense.
Thanks for the clarification. I was just thinking from the prioritization perspective of the feature request and not trying to evade your question. Yes, it is indeed possible, but my initial reaction is to consider the potential implications of OpenAI suspending your account for using a proxy. One example from Reddit:

1720214423251.webp
 
This is fantastic. Moving the spam to the approval queue creates additional work for admins, though. It would be great to have fine-grained controls added to that step. For example:

  • move message to moderation queue
  • soft-delete message
  • hard-delete message
  • delete and ban user
  • delete message, delete all user activity, and delete user
  • delete and spam-clean user
  • delete and ban user's IP
  • delete and ban user's subnet (allow exception for X, Y and Z countries)
  • delete and discourage user's IP/subnet
  • delete and ban user's email domain (allow exception for Gmail, Yahoo etc)

This has so much potential, it's insane.
 
This is fantastic. Moving the spam to the approval queue creates additional work for admins, though. It would be great to have fine-grained controls added to that step. For example:

  • move message to moderation queue
  • soft-delete message
  • hard-delete message
  • delete and ban user
  • delete message, delete all user activity, and delete user
  • delete and spam-clean user
  • delete and ban user's IP
  • delete and ban user's subnet (allow exception for X, Y and Z countries)
  • delete and discourage user's IP/subnet
  • delete and ban user's email domain (allow exception for Gmail, Yahoo etc)

This has so much potential, it's insane.
To move a message to the moderation queue, it appears that the script or the admin needs to set up a user so that the script can "report" the post on behalf of that user. If you have no problem with us creating a user such as "Forum Copilot," we can definitely update the plugin to let the admin decide if the post should be moved to the Approval Queue or simply reported to the admin.

We can also add a feature to blur the image and let the user click on it to unblur it if it is detected as an NSFW image.
 
I think the only way to take this forward and make it a super add-on is to give it the ability to soft and hard-delete, ban users, use the spam cleaner etc.
AI is still prone to hallucinating and making mistakes, so why in hell would you give it non-reversible actions such as deleting and spam cleaning?
 
AI can potentially be a great assistant for community moderation. Allowing it to destroy content and punish humans sounds a bit 1984 to me.
 
please please please someone use CGPT to design a plugin that will provide thread summaries (and deeper, thread summaries based on member or page numbers).
 
please please please someone use CGPT to design a plugin that will provide thread summaries (and deeper, thread summaries based on member or page numbers).
It is something we are working on, in addition to keep improving this plugin (Moderation Assistant). It will be similar to the AllTrail review summary:


1721407465298.webp

Please let me know if it is what you guys want? Would be nice to have a quick summary for long or important threads?
 
It is something we are working on, in addition to keep improving this plugin (Moderation Assistant). It will be similar to the AllTrail review summary:


View attachment 306560

Please let me know if it is what you guys want? Would be nice to have a quick summary for long or important threads?
Yesssss. We have auto build threads on my site that go so long and as so scattered. Being able to quickly summarize both the thread AND a specific member's post would be amazing.
 
Yesssss. We have auto build threads on my site that go so long and as so scattered. Being able to quickly summarize both the thread AND a specific member's post would be amazing.
Would you be able to PM me a few example of the threads? I can work with you to make sure it is something you want and we can go from there?
 
Back
Top Bottom