Datasets:

---

nyuuzyou

/

paperdemon-art

like 0

Provide the URLs/Titles, and/or description of the original content that you believe is contained in the dataset and infringes your copyright.

"Copyright law grants creators exclusive rights over their original works of authorship, protecting them from unauthorized copying, distribution, adaptation, and other uses. In the U.S., this protection is granted by the Copyright Act of 1976. Copyright typically lasts for the author's life plus 70 years, or longer for works created by institutions."

According to law, the entire site falls under copyright, so you are violating this law by doing unauthorized copying and distribution of artists' work who did not consent to their art and literature being scraped and used for AI. This is the reason why it is being requested that you take this down. Fair use only covers commentary, criticism, or news reporting, which is not what you are using it for.

Hopefully this clears things up, thank you.

@milkingclowns If you don't want your works to be included, list the IDs, or at least your username (even though the dataset doesn't contain it). You can also delete those and send PR.

@nyuuzyou

I have prepared a CSV file containing the URLs to take down and need a method to privately email/message you the file.

It is not appropriate for me to publish the link here publicly to my CSV data as it could be abused. It includes 49726 entries. Your dataset and my CSV also includes links that is set to only visible with direct links (meaning the artist intended to keep it somewhat private)

Please provide me with a contact email where I can email you. I need to verify it is indeed you before proceeding.

Here are some methods you can use to verify your identity

• post your email address in this thread and then email me at [email protected]
• DM me on twitter https://x.com/paperdemon from your twitter account that's linked in your profile https://x.com/nyuuzyou . I can't dm you until you follow me due to limitations of twitter.
• DM me on hugging face (i haven't been able to determine how to do this)

If you have alternatives, you're welcome to propose them.

To be clear, I am actually requesting the entire takedown of this dataset as the entirety of it includes copyrighted content that was used without consent from the artists.

I found your email address hf [at] ducks.party in another thread and have sent you an email with the CSV file.

Thank you for your cooperation in this matter.

@bogusred I think there's a misunderstanding about my request. I specifically asked for content that infringes on your personal copyright as the owner of PaperDemon.com.

I see you've sent URLs for all users' content from your platform. PaperDemon's ToS states "Users retain copyrights to all images, literature, blog posts, and content they post," I cannot remove content based solely on your request for works where you are not the copyright holder.

For content where you are not the direct copyright owner, I would need signatures or explicit authorization from the individual users who own the content in the CSV.

I'm willing to address specific content that directly infringes on your personal copyrights as the site owner, but I cannot remove parts of the dataset that don't contain content where you are the direct copyright holder.

I have sent you another email that is a CSV of some of my work. it includes 500+ works. it is not a complete list. i will need to take another pass and assemble more.

Just so you know, I'm planning to notify the artists on paperdemon about the situation so that they can file takedown requests for their works with you. It might be easier for everyone if you just remove the entire repo.

I have not received any new emails from you. Some users have already contacted me directly, like in this example, and their works are being excluded based on their username/IDs. I am committed to responding to any requests from artists about their works or other metadata regardless of the circumstances.

It is important to me to make this process as easy as possible for you and your users. Your users can reach me at the email address you already have or via Huggingface. If you're planning to make a public announcement about the dataset and ways to be excluded from it, I suggest we agree on some middle ground to simplify the process. For example, users could simply comment with a phrase like "Please exclude me" under your announcement. The CSV table you provided will also help make this process easier for me.

My apologies. I thought I had sent the email yesterday with my data but I had not. I have just sent you an email with two more CSVs that include more than 500 of my artworks to be removed. Thank you for your cooperation.

It is important to me to make this process as easy as possible for you and your users.

Taking down the entire dataset would make this very easy for everyone involved.

My apologies. I thought I had sent the email yesterday with my data but I had not. I have just sent you an email with two more CSVs that include more than 500 of my artworks to be removed. Thank you for your cooperation.

It is important to me to make this process as easy as possible for you and your users.

Taking down the entire dataset would make this very easy for everyone involved.

Or maybe I should just move the dataset to another jurisdiction :)

Seriously though, the first wave of these issues was handled almost completely automatically, including collecting work IDs even when users didn't send a CSV to the email. I'll try to fully automate this process in the near future.

Also I saw your announcement. I think it's important to tell your users that they can also use the "🚩 Report" feature here on Hugging Face. The reports are sent to the Hugging Face moderators. Users don't need to fill out and send formal DMCA notices via email, unless they prefer to.

Please provide me with an update. My works are still contained in the live dataset.

Hello,

The entries in the dataset associated with this username have been removed.

As of right now, my art is still in the dataset.

For example, art id 104 is still in images_28.zip and several more are visible on the first page of the dataset preview.

Please provide me with an update. My username is bogusred.

🚫 is only visible with a direct link

How am I supposed to identify your art if it is only available through a link? At the same time, you yourself write, "First try and request a removal with just your username" and then, in the same announcement, you publicly accuse me of not removing art that I cannot even identify this way. Therefore, you are the only one who can explain to your users why some art has not been removed - the very thing you are publicly accusing me about.

I think we're having a misunderstanding. I thought you were able to remove several others data via username. Can you please confirm the best method to have content removed. Username or urls?

I'm not sure I understand the issue with identifying the art. your manifest has art id 104 and mentions it is in images_28.zip and it's still present in there.

If you're not sure what artworks are mine, check the CSV I sent you earlier on April 18. It has links to the images and to the art pages.

I've updated the news post to revert it back to state to still use the CSV of links. If you require any further changes, please let me know.

However, I'd still like to reiterate that the correct thing to do here is to take the entire dataset down. I filed a DMCA last week but for some reason only the writing dataset was taken down and not the art dataset.

The artworks and writings contained in these two datasets were scraped from paperdemon.com. The scraping of the content was unauthorized.
The artists of these writings and artworks granted PaperDemon a license to display their artwork on paperdemon.com when they agreed to our terms of service and our end user license agreement. That license does NOT extend to huggingface, nyuuzyou or other websites. nyuuzyou and Hugging face are not authorized to redistribute these works.

PaperDemon is not a public data source. It is an art sharing site where people can display what they created as a portfolio, collaborate, and get in-game incentives for participating in challenges. It is in no way intended as a place for people to grab data or redistribute the artwork or writings.

nyuuzyou, the author of these datasets, does not have a license from the artists and authors and therefore this is an unauthorized reproduction and redistribution of protected works.

Requiring each artist and writer to opt-out is unreasonable for a dataset that was created entirely without consent. The dataset already violates the rights of the artists and Hugging Face's own content terms of service (Section Restricted Content: 4. Privacy and Intellectual Property (IP)).