00:00
00:00
Newgrounds Background Image Theme

frankoftheweek just joined the crew!

We need you on the team, too.

Support Newgrounds and get tons of perks for just $2.99!

Create a Free Account and then..

Become a Supporter!

Does Newgrounds have anti scraping TOS?

305 Views | 6 Replies
New Topic Respond to this Topic

I've been seeing allot of AI sites adding an "anti scraping" policy for their data. Which is ironic since AI models are built on theft. Does Newgrounds have an anti scraping policy? I know technically that won't stop the thieves but at least it gives artists a little more power over their creations. Plus I've been seeing how sites have some insane terms of service.

Do you guys think Newgrounds can come up with something to give artists stronger protections?


1st anti scraping on websites using AI is actually about them afraid of 'AI-loop' problem, where self-feeding harms the models. So AI sites actually want human-art sites say they are human art sites for the very reason they could avoid feedback loops. There are multitude of papers regarding this issue. Which is also why AI companies make direct contracts with website owners to get non-ai generated data for training.


Also anti scraping TOS is irrelevant as the scraping happens anyway. google for example shuts down their HTML gmail variant so they could force EVERY user into using AI and add to AI training. This explanation they pretty much gave themselves on the error page if you want to access non AI GMail.


Now if it is online, it will be scraped - as i said and I repeat myself - TOS is irrelevant*. Fact. But you can fight back against full automated scraping by using Glaze and soon also data-poison tool Nightshade.https://glaze.cs.uchicago.edu/ Safest way would be never post anything online, but that would make a lot of people isolated which isn't good. Artists should be able to interact with like-minded people.


The above project caused Adobe and few other AI companies spend MILLIONS of dollars in campaign claiming it doesn't work, so it is safe to presume it works better than we expect it does.


  • - I.e. Meta claimed you can op out, it appeared to be false and no such thing happened and they trained even when people said no. And OpenAI and MJ are known to have trained on websites even when TOs said no.

You may be right with the scraping happening anyway but I feel an extra layer of protection gives more power to the individual. For example if somebody comes up with a unique way of making art and somebody trains a model with it he won't be able to talk about it so openly about it. Plus it's an extra layer to win a legal battle and a way to demand royalties or compensation.


I know it's pretty unlikely for artists to fight a legal battle because the majority are broke but I think it's important to think about future artists and give them something to defend themselves, give them a sense of security.

We never know what people will create in the future. Maybe one becomes very successful and will have something special in his hands that rip offs might be obvious. People can help him fight. Some updated terms of service and license can serve as additional ammo.


Additionally, Tom would never sell us out. Newgrounds is the best place for artists right now, and it can get even better.


I think it's worth the shot.


At 12/18/23 02:14 AM, HenrySanchezNG wrote: You may be right with the scraping happening anyway but I feel an extra layer of protection gives more power to the individual. For example if somebody comes up with a unique way of making art and somebody trains a model with it he won't be able to talk about it so openly about it. Plus it's an extra layer to win a legal battle and a way to demand royalties or compensation.


CivitAI has communities, where people can hire other people to train models based on artist style. It is not legal, and it still happens and the host has no plans to stop this as its financially profitable. It is what it is.


There are also already AI users saying you must scrape and train before Nightshade and EU new laws get out. So if you ever heard of 'dead internet theory' it is no longer conspiracy theory, it already is today. I.e. one artist said, when she googled her own name with new Google (available in US for select accounts) - Google image search generated fake image of her style and called it hers. And then we have this 'style cannot be copyrighted' nor can 'ai works' - loop. Someone steals artists style, but then their work isn't copyrighted either and then you go read X-posts how AI generative trolls are upset someone using their prompts. Clown world.


I personally have made peace with it. I touch more grass, and do more 3D prints and props for myself and let people in fake world have their fake images and realities.


remember AI doesn't only steal from images, it steals peoples likeness, voices, music and so on. We have already had cases, where voice actors voice been used against their consent by a fan in some game mod and the VA was pretty shocked. Case of Elias Toufexis and his voice being used to add extra lines to Adam Jensen. He was not amused. Not at all. We also have people steal singers voices and making them sing whatnot or deepfakes people saying things they never did.


New common scam circulates, where people likeness and voice gets stolen from social media and then calls are made for their family to claim they need money sent to X account. or revenge ... po...eee...... adult videos are done with deepfakes because of social media has enough data to mimic someone.

Response to Does Newgrounds have anti scraping TOS? 2023-12-18 06:23:38


I don't know if it does, but I believe if they used robots.txt to block everything except the Wayback Machine, it will stop scrapers.


Although it won't stop humans from downloading the images and feeding them in manually.

Response to Does Newgrounds have anti scraping TOS? 2023-12-18 15:18:25


At 12/18/23 12:52 AM, Tenebrare wrote: 1st anti scraping on websites using AI is actually about them afraid of 'AI-loop' problem, where self-feeding harms the models. So AI sites actually want human-art sites say they are human art sites for the very reason they could avoid feedback loops. There are multitude of papers regarding this issue. Which is also why AI companies make direct contracts with website owners to get non-ai generated data for training.

Also anti scraping TOS is irrelevant as the scraping happens anyway. google for example shuts down their HTML gmail variant so they could force EVERY user into using AI and add to AI training. This explanation they pretty much gave themselves on the error page if you want to access non AI GMail.

Now if it is online, it will be scraped - as i said and I repeat myself - TOS is irrelevant*. Fact. But you can fight back against full automated scraping by using Glaze and soon also data-poison tool Nightshade.https://glaze.cs.uchicago.edu/ Safest way would be never post anything online, but that would make a lot of people isolated which isn't good. Artists should be able to interact with like-minded people.

The above project caused Adobe and few other AI companies spend MILLIONS of dollars in campaign claiming it doesn't work, so it is safe to presume it works better than we expect it does.


I was wondering why html Gmail was going away....


BBS Signature

It's all kind of out of our hands as things stand now, though things change in unexpected ways. We're in the wild west, pre-regulation period where we're getting looted with no recourse (that includes NG doing anything on our behalf), but that might not be the case forever.


NG let's you apply Creative Commons, which I know these days does about as much as a post-it note saying "don't steal" on unprotected valuables, but it might qualify you to participate in more of the kinds of class action suits that I think are coming once these things start to get audited and regulated.