#Hugging face

[redtippedfox]

So…all of my fics got scraped by that AI Hugging Face, which if your reading this Hugging Face, FUCK YOU, I spent YEARS learning to write, YEARS practicing and learning to be able to express my creativity and you just go on a cite where Millions, no, BILLIONS of writers, and artists post their dreams and stories and you think it’s okay to steal that for your dumbass AI SHIT?!

I hope your company goes bankrupt and then your hopes and dreams are crushed. I hope you never get a good day again, and I hope your wife or whatever the fuck you are divorces your ass leaving you alone forever.

As for my readers…I’m afraid I don’t know what I can do to fix this. I don’t feel safe posting my stories on AO3 anymore now that I know that nothing is being done about the issue. I may just post them on here…but for now I don’t know what the future holds.

To all my artists and writers…do not let them win. AI will never win, no matter what. I know right now it’s seems hopeless but if we keep fighting then there is hope.

No machine can do what we do so keep your head high.

To Hugging Face.

FUCK YOU

To AI lovers

FUCK YOU

[trypo-p]

More info here

[indefenseofmasochism]

aight

huggingface ao3 scrapers try not to create a false narrative challenge level impossible. try not to commit hasty generalization challenge. try to touch grass challenge.

here’s the thread if yall wanna reblog with the logical fallacies yall find in their arguments.

[huggaboos]

Attention to all AO3 Writers!

I do not normally make announcements like this, but I just found out this morning that 12.6 million works from the AO3 and other sites were scrapped to train an A.I. Model called 'HuggingFace'. Ao3 had nothing to do with this, so I would check to see if your works have been copied/stolen. The easiest way to check if yours was stolen is to look at your public works ID. It apparently had taken works posted before March of 2025 this year.

The model took ID's ranging from 1 - 63.2 million, so I would check any recent ones that you made.

[chrissypoothepanda]

It's too late to stop AI.

It's too late to stop people from using it for stupid asinine bullshit.

It's too late to convince people that the rampant proliferation of data centres are actively harming our environment.

It's too late to convince people it's use is going to lead to a greater chasm between those who have in excess and those that go without.

It's too late to deny it's usefulness as a tool for learning, for scientific research, for convenience.

It's too late to walk back the last 5 years and try to start fresh with our current hindsight.

But we can insure ourselves for the future.

We can implement legislation that insists AI generated or enhanced image and video content is flagged as such.

We can implement licences for it's commercial use to protect the livelihoods and interests of those who's professions it would otherwise wholly absorb. Regulations for the companies that choose to use it as an alternative source of human labour purely for the sake of the bottom line.

We should be putting warning labels on it like we do cigarettes, alcohol, medication, weapons, and heavy machinery. Those things are all harmful if misused or engaged with irresponsibly, but we have learned to adapt to life with them by adapting and enforcing rules around their use and access.

[oceangirl24]

reading about the bastard nyuuzyou and their counter arguments and realizing that sometimes it's valid to wish death and harm on other people :D

I'm going to be completely honest here, I disagree.

It is frustrating, heartbreaking, and disgusting the attitudes that are on display by the pro-AI crowd on hugging face.

The worst parties here are actually not nyuuzyou (this is not a defense, what they did was and is detestable). There are users on hugging face actively attempting to use the dataset to harm others as a power flex.

However, both sides have acted in ways that are appalling.

One so-called "anti-AI, pro-AO3 author" went so far as to dox nyuuzyou and comb through their personal social media and posting what was found while threatening to find and release more. Their remarks about the social media content they reposted contained some of the vilest transphobic rhetoric I've had the misfortune of reading.

I have no idea if they really did dox nyuuzyou or if they were who they claimed to be. They could have easily take content from an unrelated user and posed as an AO3 defender to make the rest of us look bad.

I say because on the other side a "pro-AI, your work is mine for the taking" user who claims to be using the AO3 data in training right now and who posted an image of the alledge training, also launched a vile, transphobic attack on an author whose work was stolen because their name contained "authoress".

Different people or the same person? Doesn't matter. Wrong is wrong.

Thankfully, the huggyface mods did do something right and deleted the gross content and locked the thread.

Look, I was among those whose works were stolen. My art. My original character who I've developed since I was eleven was stolen. I just fought off a harrasser and plagiarist in my own fandom and now this.

Yes, I'm angry. Yes, I'm devastated. But I'm not going to wish harm on anyone either.

It won't change anything.

If you've been following this then you know there is a user claiming to be legal counsel and threatening those who file DMCAs even on their original works with lawsuits and says AO3 will be targeted as well if they pursue things legally.

A lot of this is rage bait and scare tactics. While there may some truth admist the name-calling and threats, OTW exists for a reason and if there was nothing legally they could do without getting all of fandom in trouble, we'd know something by now. But we don't.

The only thing I can say is be patient and wait to see what AO3 says.

And please don't engage in anything that can be construed as inciting violence, harassment, or doxxing.

Please.

[demonicintegrity]

Hearing that public works on ao3 were scrapped for Huggingface is so infuriating. Scrapping in general is just both a morally reprehensible way to gather data, and stupid too. If you're really going to build datasets, having a narrower parameter then "literally everything" is the very start.

According to this article it used to be a chatbot but moved onto "primarily used for natural language processing tasks, such as text generation, sentiment analysis, and chatbot development." At a glance, I would not be scrapping a huge fandom archive for that sorta thing, considering not-insignificant number of fics that aren't written with the best grammar and also crackfics.

Going to the website itself and scrolling down I see this

I am not tech savvy enough to have extended commentary on the models or datasets outside of the fact that if this company wanted to be a hub for legitimate and ethically ai development, you think they would put more care into how they were getting their information. It's those spaces i'm more concerned about.

Now, it seems than they're uploaded by different people and not just Hugging face itself, but I digress. A skim has already shown different dialogue and character creation related bots made to generate.

Here is the post of the user on hugging face who made and uploaded at dataset of ao3 fics. Needless to say, do not make accounts just to make threats and drag this creator. Knowing its a user and not the action of the website itself is only comforting until you wonder why/when will the website screen the datasets being uploaded to make sure they're ethically sourced. (This also doesn't appear to be the only dataset made with fanfiction, but by far the largest.) It appears that Hugging Face is doing what they're legally obligated to do.

Access to it has been restricted thankfully. Following that link leads me to this

and the comments are as ugly as you can expect them to be. (No, fanfiction isn't theft. It is protected under parody and fair use laws. Regardless, i cannot fathom why you would want to use them anyways.)

If you're curious if it's been scrapped, this tiktok said any work with the work id under 63,200,000 appears to be scrapped. Basically anything made mid this march and before. All of my fics, including my one original work has been scrapped. Since ao3 lawyers are on the case, there's nothing anyone else can do now but wait. There is no point in submitting your own DMCA claim because Ao3's covers all of us.

That being said, it's irritable and disheartening. This is not in the spirit of fanfiction, this is not legally or morally ethical, and it's just irritating. None of these works had consent to be reuploaded or used in this way (no matter how those ai idiots think things work. Just because it's on the internet does not mean it's free use)

The conversation now is how to prevent this from happening again. Many users are setting their work to be privated, so that way only those with an account can read fics. I fear that is a very bandaid on the bullet wound type solution, because there's nothing stopping someone from making an account and then scrapping. Ao3 themselves have their own methods to try and counteract such things. And largely, we will be relaying on the team to handle such things. (This is why that donation drive is important, i can imagine the poor legal department is going to be clocking a lot of overtime soon.) I suppose another method is just to go old school. Printed zines for yourself and your circle of friends and not online I guess.

What now? Not sure. Obviously arguing with ai bros on their own turf isn't doing anything constructive, not that I really expected anything else from the general fandom nerds tbh. It's just. A super fucking frustrating reality of the internet now. I can only hope that when enough idiots do something like this and face repercussions, it'll be a widespread lesson on Ethics in Business and 101 for all.

It sucks, because back before "AI" was associated with chatGPT and datascrapping, it was understood to be broad range of tech and code that can be really good and helpful. Unfortunately, aibros of today do not know how to, nor will put in the hard work, to making something more constructive than ArtTheftArtGeneratornumber3000 and ChatBotThatSucks2000. With so much of the slop, it's taking away time, energy, and resources from people trying to development meaningful AIs to recognize cancer and process information humans can't in a timely manner.

Edit: Its worth knowing this is not the first, nor likely the last time, ao3 has been scrapped. This isn't new but it's getting annoyingly common. And this particular dataset is *huge*

[wauln]

[ DeepSeek AI Exposes America ]

[jamalir]

Open-R1: a fully open reproduction of DeepSeek-R1

[2pdreamtale]

I made a big AU and I hope you like it 😘 But my English is very bad, so if there is anything wrong, you can ask me some questions. The name of the story is "Opposite Dream" or "2p Dreamtale" All here The characters have all become 2p. It is a similar AU to Dreamswap, except that many of the original characters have become 2p versions (^-^)

[peterbordes]

Groq just made Hugging Face way faster — and it’s coming for AWS and Google.

Hugging Face + Groq = Game Changer for Builders

Groq is now integrated into the Hugging Face Playground and API, unlocking the speed to bring your most ambitious ideas to life — faster than ever.

From smarter agents to real-time copilots, the future is yours to build.

#AI #artificialintelligence #huggingface

[the-ai-perspective]

Hugging Face's Open Computer Agent Debut [Free Episode]:

[mistytheghost]

so, all my fics in ao3 have been scraped into that AI dataset.

as if i needed any more deterrents from posting… i think from now on i’ll just stick to sharing links to my docs in private.

[oceangirl24]

Locking Fics to AO3 users only.

I really hate to do this, but until the nuyuuzyou situation is resolved my works on AO3 will be locked to registered users only. I am really sorry.

Guest readers: my works are still available to read without logging in on FFN and Wattpad. See pinned post for links.

After reading this person's (and others like him) response to the theft, I just can't leave my works public right now. The attitude being displayed is "if it's online I have the right to take it and use it as I please". One such user even said they wanted to feed AO3 works to AI to produce "interactive" fanfiction to their taste. Another user threatened legal action against everyone filing a DMCA request, including OTW, saying to file "false" takedowns is illegal and subject to lawsuit. 🙄

It is interesting that the discussion the above remarks were made in has since been deleted by huggyface. In fact, a lot of the comments that were in these discussion threads have been subjected to mod intervention.

You can read what's still available on huggyface:

Nyuuzyou

Paperdemon-art discussion

AO3 data set

Just a little disclaimer- as tempting as it is, please don't harass the user. As gross as what they've done is, it helps no one and you sink to their level which is exactly what they want.

Take the high road but stay informed.

Love you guys. Readers, creators, take care.

[transfaerie]

Continuing to work on the Hugging Face agent's course on stream. Today: Unit 2, Agentic Frameworks

twitch.tv/transfaeries