DarkOverord

Oh no a Hedgehog.

DarkOverord's 88x31 Web Button Lol why isn't it centred?

Hi! I'm DarkOverord! I'm an agender furry who also does digital art, and it's almost entirely furry.

🦔
Usually I'm a vampire hedgehog, sometimes I'm not a hedgehog but still ΘΔ

🎨
Art only page over at @DamnRedDragon

🎮
Generally expect most fandom related stuff to be Furry, Pokémon, Touhou and Sonic.

Character References
Icon by bobateabandit
Header by Kiryu


My Last.fm


staff
@staff

along with a lot of other AI art announcements we're much less pleased about, deviantart announced this morning that they have created a standard for flagging works of art that you don't want to be bulk-scraped for machine learning datasets. in response to this, we are adding these flags to every page on cohost.org. we're off today for a federal holiday, but we punched in for five minutes to get this done because it was easy to do.

we have no plans to start offering any AI art generation tools, and we also have no plans to allow you to opt into having your art scraped. if you want to opt in, you're welcome to post your art to another site as well. (and, unfortunately, even if you don't want to opt in, nothing protects you from people reposting your art to other sites which don't set this flag, which we all know happens constantly.)

it's possible that dataset or model vendors will see platforms blanket-tagging pages with noai and noimageai, and decide that the original intent of the standard -- to reflect an artist's personal, conscious decision not to have their artwork scraped -- has been compromised1, and their scrapers will start disregarding the flags altogether. in this case, they'll probably blame us for not playing fair; so be it. at least then they'll have to admit how thin their commitment to respecting this standard was in the first place.


  1. indeed, this has happened before, with the W3C Do Not Track standard to allow people to opt out of tracking cookies -- Microsoft began setting this flag by default in Internet Explorer 10 and the internet advertising industry quickly responded by announcing that they were respecting user agency by ignoring it.


You must log in to comment.

in reply to @staff's post:

very nice! glad to see you guys take the first steps i've ever seen against this type of exploitation - have you considered putting it in your TOS, while you're making changes to that too? i have to imagine that's going to be the next step for sites to try to stop abuse.

it's already in there! (even the current version -- new version still coming early next week!)

  1. Use Restrictions. You agree not to rent, retransmit, disclose, publish, sell, assign, lease, sublicense, market or transfer the Services or any portion thereof (including our Content) or use it in any manner not expressly authorized by these Terms. [...] Without limiting the generality of the foregoing, you understand and agree that you will not: [...] Download, scrape, post, or transmit, in any form or by any means, any part of the Services.

as some folks said in our defense in the ToS debates, most of the meaning of the Terms is in how we enforce them, and if we have evidence that LAION/StabilityAI (or anyone else building an ML model) is scraping our site in spite of this clause, we won't rule out legal action.

this is of course great and thank you but I daresay AI people don't really care about what artists say or think or ask or tag, they'll just scrape regardless, they don't care whatsoever. But I'm glad for this, it's the most anyone can really do