This is not looking great for you
First Amendment & defamation lawyer. Now: Lead Counsel for Tech Policy at @thefireorg.bsky.social Former: Free Speech Counsel at @TechFreedom.org Illini/music junkie/oofnik. “A snarky gay lawyer Jessica Fletcher.” https://linktr.ee/aricohn
46,970 followers 781 following 8,410 posts
view profile on Bluesky
This is not looking great for you
Yep, can't fault the party with all the risk yet still no ability to fight against obvious bad-faith notices
I'd certainly call deindexing sufficient to support nominal damages 😈
Not for nothing, but the NO FAKES Act will make this problem worse. It incentivizes using tools *exactly* like TakedownsAI for preemptive content removal.
(but yes, your idea would do a lot more justice in terms of deterrence for the folks who don't have the knowledge/money to engage with the system, so I heartily endorse it)
Kinda thinking about setting up a little side hustle of going after TakedownsAI and the creators who use it. Maybe a little legwork, but they've already made the cases viable. Would seem to be a shame not to take their money.
"We would not file the bad-faith notices, but that would be less money for us so we do it anyway" is quite a thing to say on the record!
Now, now
Put another way: failure to communicate the message has not been sufficient to strip expression of protection
It doesn't matter omif the answer is ultimately useful or not. The desire that it be is enough.
I don't think "protection depends on what the ideas are" is going to be a winning argument. Besides this is all undercut by the desire to impose liability because they advertise a "truth machine." "providing useful answers to your prompts" is expressive intent even if an answer isn't useful.
By the way, while Kunal's LinkedIn profile says he's in India, the company TakedownsAI is under, "Venus Tech" lists its HQ as being in WY. So probably not so difficult to go after these bums.
If I had to guess, it boils down to the recency of acceptability. The former they grew up able to use, while the latter has been taboo since they were born.
These guys still haven't passed shapes and colors, you want to have expectations about *math*?
These people are unrepentant racists but also too cowardly to use the slur they so desperately want to.
No it suggests that if its a specialized technical tool, that's the case
Sure sounds like Kunal Anand is admitting to filing bad-faith DMCA takedown notices here. www.404media.co/how-onlyfans...
From Winter v. GP Putnam, here are cites to some of the existing cases at that time; there are further ones but the reasoning of the 9th here makes there relevance explicit so wanted to provide that.
I don't think that's right. If true, is the parade not expressive if the organizers try to exclude certain messages but fail, resulting in an unintended viewpoint being disseminated? The fact that they are trying to eliminate the output in the first place means the result is inherently expressive.
But if there are ideas being conveyed, the entity organizing, creating, curating, whatever them has expressive interests. That can't be obviated by saying the ideas weren't "speech" because they were created by some intermediary without speech rights.
In fact that's an argument I believe that was advanced below. And that's the argument TX/FL advanced to try to claim their content moderation laws didn't violate the First Amendment: that there's no speech by the platforms themselves.
Right, but that gets down to the point that there was an idea being conveyed that the organizer didn't like. Whether or not it was "speech" was not relevant. What mattered was that there were ideas that the organizers disapproved of. Nobody claimed the marchers' speech was the organizers'.
Your pushback on Hurley puts the locus of analysis in the wrong place. Hurley didn't depend on whether the participants were engaging in speech. What it said was that the *organizer* has expressive interests. Whether or not the participants were engaging in speech of their own was immaterial.
Wanted to come back to this now that I'm not multitasking and on my phone. I appreciate that you took the time to go read and digest the case and construct an argument. That's more than most folks will put into it so I wanted to specifically call that out. However (cont.)
What's good for the goose is good for the gander. I think there are some interesting questions about how ordinary liability would be analyzed in terms of assessing mental state, but I think conceivably an AI company can be held liable for defamation.
I say that not to be glib, but to underscore that the impetus to use the law to ensure safety often leads to shit results, but still no safety
Life is not a wholly safe endeavor! The answer isn't safetyism. That's what leads to CPS taking kids away because parents let their kid play in the park alone.
You might argue that, but you're wrong
I'll try that regimen next time!
I've given up on getting the sage to crisp. It happens like 1/30 times, even when the browned butrer otherwise turns out great.
It is not authorless. The creators of the AI have many expressive inputs into the system that are directly intended to shape the message. Changing the inputs changes the outputs.
You're engaged in circular reasoning
Love it, they've got a new supporter
I mean wouldn't you fire the guy who believes in the thing that's critical to your case? It's what all the smart clients are doing lol
If you want to talk about question begging, this is actually it. Your rationale presumes the conclusion.
This is sophistry. The entire argument here is that it's not speech *because* it wasn't selected or controlled. Thats the only way to say that it isn't a human creating it.
It's almost like people want an excuse to do nothing. It's like the legal nihilism of "well the Supreme Court has been bought off so the law means nothing anymore."
It's become more and more clear how so many people were able to be convinced last time around that the election was "stolen." There needs to be a renaissance of civics education in this country. There should be years of mandatory instruction on it at each level of education.
The state of civics knowledge in this country is a national crisis. Seriously, so much of the doomer bullshit here comes from people who do not know, at a basic, mechanical, who does the work and where level, how America works.
She the harm, and the reason it is dangerous in the first place, is because of words and ideas, it's entirely circular
Again, that's not true. Whether a duty can be applied in the first place is a matter of law.
Also, if you're saying the harm is not mitigating the risk, then the speech integral rule doesn't even apply! There's no such rule as speech integral to the risk of unlawful activity happening. If nobody ever got hurt because the risk never materialized, would this still be speech integral?
Fraud involves unprotected speech. That's an entirely separate category of speech! Here, you're saying that the speech is integral to unlawful conduct, which is providing the harmful speech itself. That's circular.
I know what medmal is. The giving of professional advice is subject to more regulation (and thus more liability), in part via a recognized special relationship. There is no special relationship here.
If you want to say there's a false advertising claim, maybe that works if there's a piece of commercial speech. But this is not a good fit for products liability, IMO. I think people are more troubled by the idea that there might not be liability here than they are with the impacts of that liability
And you can say it's about things other than speech, but those things all relate to the fact that the AI outputted harmful content. There's no magic wand that can make that fact go away.
Professional speech is subject to more regulation, so that's not a good analogy
(I won't claim exactly encyclopedic knowledge, but I've read most of the products liability casealw that atrenpts to extend it like this)
But here you're not talking about bad instructions to the user. (And again, the only instruction manual case I know of is NOT products liability)
Liability for failure to mitigate risk because of words is famously prohibited. There's no reason why that doesn't apply to the risk of the words appearing in the first place
You're looking at a false advertising claim at most there. And that's a commercial speech case.
Is there a case where "speech integral to products liability" was held unprotected? I can't recall one. It's twisting doctrine because it's entirely circular. It declares speech unprotected because it is unprotected. Because again, liability hinges on the speech.
The code did not, independently of content, generate an unsafe condition.
I think that's a completely artificial distinction. If you take away the content, there is no foreseeable harm.
Either way, I suspect we're entering the "do not attempt to stop chainsaw blade with genitals" era of AI.
I just don't think twisting doctrine to assuage moral panic is going to lead us to good places. Yes, this was tragic. But we've seen before over and over that the rush to make sure the law allows "accountability" in response to emotionally difficult circumstances usually backfires.
And that's not the speech integral rule, that would be quite an expansion of it.
So are all the woo woo books about how rocks can cure your cancer or whatever unprotected?
That's a twisting of the speech integral to criminal conduct rule
But the reasonably foreseeable harm IS the content! You just can't separate them.
Though the latter would not be surprise using Grok
Slurs not sluts lol
I think the artificial distinction between the code and the output also raises another problem for this theory. If the code is expressive such that the government couldn't say "AI must output racial sluts," then liability for code that doesn't prevent recommending suicide is obviously a 1A issue.
Literally in the history of ever, nobody has said that ChatGPT has First Amendment rights. You might try having a less simplistic approach to the question before saying something this silly.
I think that's a tenuous line
Again, there's no requirement that a speaker intend something with this level of specificity. If solving the problem requires changing expressive inputs (coding/training, which even you acknowledge is expressive), the output is expressive. That's Hurley.
You simply can't get to the harm without the content of the output. And while I understand it's frustrating to think that maybe nobody can be held liable under doctrine as it stands, I think that the reliance on output not being speech at all indicates that this theory is an end-run
1):There is no requirement that the words or ideas need to be expressly intended for something to be speech. That idea has been explicitly rejected on multiple occasions. 2) The code itself did not cause harm; it didn't physically kill someone. The code resulted in the expression of harmful ideas.
Sure there are, it's in the aeronautical charts line of cases
I mean, are you surprised
If it were me I'd tell them to kiss my ass and see if they want to dance (issue a subpoena)
Got that right
Not at this point, though the committee could issue a subpoena if they don't at which point there is one
Although I'm not so sure (sadly) that we want to encourage these particular people to be more literate
As in the response to this deeply deranged letter, I assume (judging based on your background!)
That's actually very good to know
This is super cool and I did not know this somehow, so thank you!
Oh sorry wrong list item. On this one yes, bad writing. They want information on coordinated efforts to influence content, which was later found to violate policy
Yes they are asking for that. Presumably to assess whether or not they think Wikipedia is doing a good job of it (also none of their fucking business)
7/ What we said then, remains true now: Congressional investigations into protected speech, based on viewpoints that Congress disagrees with, trying to find some unlawful activity to punish evokes some of the darkest days for free speech in U.S. history — ones we shouldn't repeat.
6/ Unfortunately, this is part of an ongoing trend. Last summer, FIRE joined a coalition condemning the committee's aggressive demand for the private correspondence of American Muslims for Palestine related to Hamas' October 7, 2023 attack against Israel www.thefire.org/news/house-o...
5/ That is deeply chilling, and this is all more than a little galling given the (justified in principle, if not in fact) Republican outrage in recent years about allegations that the Biden administration pressured social media companies to take down protected speech. Jawboning is bad, period.
4/ To help it investigate this unambiguously protected speech, the Chairman Comer wants Wikimedia to hand over not only information about editorial policies and how it has resolved content disputes, but **also the identities of those who engaged in disfavored speech.**
3/ Wikipedia may aspire to provide a platform with “a neutral point of view” that avoids “bias,” but it is under no obligation to do so; its editorial policies are protected by the First Amendment — and none of Congress’ business.
2/ The committee doesn't hide that it is targeting viewpoints with which it disagrees, citing efforts to “inject bias into important and sensitive topics,” & “advance antisemitic and anti-Israel information." Otherwise known as protected speech. Letter: oversight.house.gov/wp-content/u...
1/ This week, House Oversight Committee Chair James Comer demanded that the Wikimedia Foundation hand over information on editorial policies & individual contributors to help the committee investigate "propaganda" on Wikipedia. These demands are offensive to 1st Amendment values
Honestly it was already pretty much the best state fair, this is just overkill
Minnesota state fair free press booth doing the heavy lifting @aricohn.com
lol
Praised and promoted it just generally?
It's a hell naw from me dawg
(and also, when used as a replacement, is likely to be absolute shit at it)
Let's be clear: zero argument from me insofar as that any app that explicitly purports to practice a licensed profession in which there is a legally-defined duty between provider and providee (doctor, lawyer, etc) is regulable to the same extent as a human provides those professional services!
I do think that they are unconstitutional. And to the extent that states argue they are constitutional because suicide is a crime so it's incitement, I think that's reprehensible. I want no part in validating a law that says if you fail in a suicide attempt, you should have a criminal record.
I DO however, think that saying people should use AI (even purpose-built) for therapy is among the stupidest, most hubristic, emblematic of tech co excess/malignity, things one could possibly expel out of their word hole. So we agree at least on that.
That runs into the same problem as the "profit motive = commercial speech" argument, IMO. I don't know that "answer machine" is a claim that can be enforced in that way. Can we sue an encyclopedia publisher if articles are wrong + cause some harm? Doubtful.
oh good sam alito had an *idea*
But I think the strict liability point is exactly why the courts hew so far away from applying to harms from words/ideas
ah ok, could have been me that misread
(but I still love you dearly)