THE SCRATCH TEAM HAS ADDRESSED THE SITUATION, SO THIS POST IS NOW OUTDATED.

Moved to the “Show and Tell” section.

Project link: https://scratch.mit.edu/projects/1270059283/

Hi here! Today I'll talk about the current Scratch situation, so let's get started!

As you might know… Scratch just updated their Terms Of Service (TOS). At first place, Scratchers don't think it's a big deal, so they don't need to read them. However, when they actually read the TOS, they found out that TOS said that “User Content you submit to the Service to create AI-generated outputs is referred to as “Input””. Which means that anything you created (and/or shared publicly on Scratch) will be used to TRAIN AI, and the “AI”, they meant their up-coming Creative Learning Assiant (CLA) on Scratch. Keep in mind that “User Content” means contents of an user on Scratch, including projects, messages, comments, etc.

But this section (section 8.9) is where it gets EVEN WORSE, because they also said that by using the AI, you give Scratch permission to… Use it to improve Scratch, for commercial purposes, train AI models and let third parties use it (for free or paid). The scariest part is "improving AI models and allowing third parties to do so”. Because that line means… If you use our AI tools, they can use your AI interactions to make AI better, and they may work with partners to do that, so they might even collab with other companies to use your prompts sent to the AI for training the AI.

And as you expected… A lot of Scratchers went crazily mad about this update, most of the recent posts in the “Questions About Scratch” section of Discussion Forum on Scratch are just complaining about it, some of them are so mad that they even hated and roasted Scratch and the Scratch Team about it, which is… extremely bad for an educational kids programming language like Scratch.

However, the elephant in the room that a few people talked about… Is how “weird” is the new TOS of Scratch. It's very vague, broad and corporate, literally just like Roblox, so that's why it sounds scary. I can't even understand what are they trying to say.

So, what solutions for users and staff to end this chaos?

Well… For users, they should not:
- Scold or curse Scratch, the Scratch Team and the Scratch Foundation.
- Start a Flame War on Scratch (like on the profiles of the ST members).
- Raid or harass the members of the Scratch Team or the Scratch staff.

But they should:
- Convince or encourage them to re-change or fix this update.
- Consider them to train AI in a way that the community accepts.
- Contact them to talk about this situation via “Contact Us” button.

For the Scratch Team and the Scratch staff, they should:
- Put the AI features under opt-in AI Terms.
- Add an option for users about using User Contents to train AI or not.
- Add a FAQ that is about the new update so the users can ask about it.

Summarily: Scratch needs to replace ambiguity with guarantees. Or… Scratch should stop being corporate like Roblox.

That's all I want to talk about the situation, remember that everything I said in this project are mostly from myself, I might be wrong sometimes.

For the Scratch Team, I hope you can address this situation and find a way to end this situation and solve the big problems, thank you for reading this.

Please don’t make guides. This fills up QAS.
If you’d like to make guides, go to the Scratch Wiki.
Please use contact us for rants.

cookedasparagus8 wrote:

Please don’t make guides. This fills up QAS.
If you’d like to make guides, go to the Scratch Wiki.
Please use contact us for rants.

How about other people in the section? They even cursed Scratch…
But oh well, let's not talk about it since this post is now moved.

UPDATE: After reading the new post from the Scratch Team about the situation ( https://www.scratchfoundation.org/ai ), I can finally take a breath. The Scratch Team is NOT using our data to train AI anymore, and also the data that Scratch will collect are NOT our personal information. They’re listening to community concerns and invite feedback as they work on AI features. Phew…