Imagine a vibrant community like Scratch, where kids and adults alike are free to explore their coding creativity without worrying about stumbling upon inappropriate content. By integrating an AI detector to identify and remove such content, Scratch can maintain a fun, educational environment for all users. These AI-powered detectors efficiently scan projects and comments, ensuring everything aligns with community standards. This allows users to focus on developing their skills, collaborating with others, and enjoying their time on the platform. The addition of an inappropriate project and comments detector also adds an extra layer of safety, making Scratch even more secure. This system is purely functional and not designed to engage in conversations like a chatbot. Without these safeguards, the platform risks being blocked or banned in various places, which could significantly limit its accessibility and impact. This extra measure ensures that Scratch continues to be a trusted and beloved platform for users of all ages.

As I have posted on similar suggestions to this:

AI isn't perfect, not by a long shot. Using it as the only thing to detect if a project is appropriate or not will wind up with perfectly appropriate projects being taken down while inappropriate projects stay up.

As a kind of early warning system, if could work, but as long as it can only give at most a warning to the project creator that their project may or may not be inappropriate, but still allows the project creator to share and modify the project.

what if I would send a notification to ST for review if it unsure it could also check projects for copyright music or inappropriate and comments can be similar to Roblox but instead it will not let the comment be shared

I'm going to assume you used AI to write this (apologies if I'm wrong, it just looks very similar to AI's writing style and Grammarly also flagged this as AI)
If so, then maybe you should let the AI know that there already is a filterbot in place. While this filter doesn't use AI, it has a blacklist including many phrases against the community guidelines, including censoring ways people dodge the filter (e.g. unicode) and it also works in multiple langauges. With this considered, I don't see the need to implement AI, especially when there are times it can slip up. In the case that a comment does make it past the filter, you can report it.

I won't trust AI to do content moderation - or anything else on Scratch for that matter - until it's more than a bloated and glorified autocomplete. If you can't trust it to count letters in a word, why would you trust it to moderate a website with rules as tight as Scratch's?

(I know not all AI is gpt-based, but I don't think non-gpt-based AI exists for moderation yet)

8xa wrote:

I'm going to assume you used AI to write this (apologies if I'm wrong, it just looks very similar to AI's writing style and Grammarly also flagged this as AI)

GZ says 100%, though it isn't totally accurate.

OP, if you did use AI, be aware it is not allowed on the forums afaik.

Next time actually use your fingers to type something instead of using ai for your suggestion

if it was roblox moderation (uses ai) then users would be banned for saying “hi”

createrpro103 wrote:

Imagine a vibrant community like Scratch, where kids and adults alike are free to explore their coding creativity without worrying about stumbling upon inappropriate content. By integrating an AI detector to identify and remove such content, Scratch can maintain a fun, educational environment for all users. These AI-powered detectors efficiently scan projects and comments, ensuring everything aligns with community standards. This allows users to focus on developing their skills, collaborating with others, and enjoying their time on the platform. The addition of an inappropriate project and comments detector also adds an extra layer of safety, making Scratch even more secure. This system is purely functional and not designed to engage in conversations like a chatbot. Without these safeguards, the platform risks being blocked or banned in various places, which could significantly limit its accessibility and impact. This extra measure ensures that Scratch continues to be a trusted and beloved platform for users of all ages.

Your PFP is made by ai
This was made by ai.
Moderation can NOT be made with ai.

This is a duplicate of an earlier suggestion.
Please use the linked topic to keep the discussion all in one place

For future reference, you can use the Suggestions Directory or Ocular to check for duplicates.

It looks like this is a duplicate topic of this one over here so I'll close it to keep the conversation all in one place.

Please use the existing topic in the link above.