AI companies copy all written works they can get their hands on and call it fair use, if someone does it to their models it suddenly becomes "unauthorized distillation" and should be actionable in court.

The double-standard is ridiculous.

https://www.theregister.com/2026/02/14/ai_risk_distillation_attacks/

How AI could eat itself: Competitors can probe models to steal their secrets and clone them

: Just ask DeepSeek

The Register

@seanfobbe

The fact that Google put their own needs front and center in that threat intel report instead of those of their customers and the public was frankly off-putting.

@seanfobbe NOTHING good will come from AI. Hear me now, quote me later.
@seanfobbe @Ponygirl
A lot of billionaires might go broke…?
@dgar @seanfobbe You got me there 🙄
@seanfobbe They should chill. it's fair use, remember? 🤣
@seanfobbe "They are stealing the stuff WE stole!"
@seanfobbe We're "training on" their reasoning; it's definitely transformative fair use!
@seanfobbe What exactly do they mean by API router loopholes?
@doctorlaura @seanfobbe no clue but I would assume they mean a distributed api router amongst all the providers. It reads like OpenAI wants a government mandated single api router so a distillation attack can’t be spread across multiple providers. I’m guessing openAI thinks they should manage and get access to the data in the single gov mandated API router.
@seanfobbe not surprising and as @nicklockwood posted earlier (I’ll have to find the post and update here), and I’m paraphrasing, that the law will always side with to corpo-fascist because they are dictating the laws not the publicly elected criminals, and certainly not common sense.
@seanfobbe It stuns me that anyone with proprietary software would allow it to be uploaded to somebody's AI model for the purpose of improving, refactoring, or "finding bugs".
You've just given away your most sensitive IP, to somebody who did not sign your NDA. In my work, previously that would have been unthinkable.
@seanfobbe disappointed the reg didn't get comments from the companies on the obvious double standard here. Or at the very least point it out 😞
@deliverator Yep, I find this disappointing as well. The Reg article is quite one-sided and doesn't address the fair use double standard at all. They usually do much better...

@seanfobbe I knew it would take a situation like this for the techbros to actually consider pushing legislation.

The whole "nuuuu, legislation will stifle innovation of a vulnerable new industry uwu" to getting lawmakers involved once other foreign competitors start stealing the datasets they stole in the first place is peak capitalist greed and systematic corruption.

What would be hilarious is if the legislators say "our hands are tied due to the 10 year ban on AI regulation provision in the big "beautiful" bill, leaving these techbros high and dry to fix this on their own, but I don't think they realize the shitstorm that would ensue if they ask the dementia-ridded orange for favors to skirt around said law in this instance.

@seanfobbe @michaelgraaf […] #OpenAI has to be the most insufferable company in the world. They can steal from the whole world and guzzle all possible resources. But no one can give them a taste of their own medicine even a little bit.

https://dair-community.social/@timnitGebru/113909909764399262

#theaicon #llm #deepseek

Timnit Gebru (she/her). (@[email protected])

OpenAI has to be the most insufferable company in the world. They can steal from the whole world and guzzle all possible resources. But no one can give them a taste of their own medicine even a little bit. How long till they use this to say give us even more resources and this is why we can't release anything. PS: Didn't OpenAI say if a competitor achieves "beneficial AGI" that they'll consider their mission accomplished 😂 https://www.ft.com/content/a0dfedd1-5255-4fa9-8ccc-1fe01de87ea6

Distributed AI Research Community
@seanfobbe Chinese AI Labs being incredibly BASED over and over again

@seanfobbe the double-standard needs to be outlawed.

  • Just like "#AI" content isn't copyrightable, neither should it's use be restrict-able by the companies that build that shit!

#sarcasm #commentary #AIslop #Copyright #Copywrong

@seanfobbe "Democratic AI" feels coming out straight from a parody.
@seanfobbe Oh no! Theft machines are stealing from the theft machines?!?! The HORROR!