Florida sues OpenAI and Sam Altman over AI risks
politico.com
politico.com
It is probably not true that ChatGPT has resulted in an increase in murders and suicides, and certainly it would be very difficult to prove liability on OpenAI for this. It reminds me of the campaign in the 90s against video game manufacturers for "corrupting the youth".
But I also don't think they expect to win. They just want to show that they're doing something to fight tech companies and AI.
[1] https://www.cbc.ca/news/canada/british-columbia/sam-altman-t...
Just to clarify for anyone not paying attention -- Anthropic has written postmortems detailing their Claude Code monitoring and how they "coordinated with authorities" as they "gathered actionable intelligence" from users creating bad content [0].
Because it's obvious? Any interaction on the cloud = someone else's computer has 0 expectation of privacy. Are we next gonna pretend being shocked that google queries aren't private.
There's no legal privacy such as doctor-patient confidentiality or lawyer-client confidentiality. It would not surprise me that people want some sort of guardrails protecting the public from recipes-to-do-evil. Based on how much pushback the AI industry has been doing in response to guardrails about hate speech or porn, I expect this to be fought to the death by the AI industry.
Either one seems so glaringly artificial and transactional it'd be more depressing than loneliness.
The government did intervene though. They threatened to regulate the industry if the industry didn't regulate itself. So some/all the big industry players got together and created their own independent age rating agency that they all agreed to use.
Whoever was suing won in the outcomes department.
Politicians in general have a bad habit of threatening and passing speech laws that judges torch on sight.
The reputation of Florida comes from having a very broad public records law that requires publication of numerous details of police reports that in other states is kept confidential. That means that sensational stories are much more likely to make it into the news: https://digitalcommons.wcl.american.edu/facsch_lawrev/2117/
>The problem with this theory is that it incorrectly implies that Florida's Public Records Law offers journalists advantages in writing stories that other states' laws do not. Despite the broad grant of access to police documents that Florida'sopen records law provides, other states' open records laws similarly provide the public with access to arrest records, incident reports, and, although to a lesser extent, mugshots. Other provisions of Florida'sPublic Records Law that contribute to the ease of access to Florida'spublic records compared with other states equivalent laws are largely irrelevant to FloridaMan's existence. Even coupled with the characteristics of Floridaandits residents that many people claim are unique, the open records law-based theory for Florida Man's existence falls short of explaining the phenomenon.
My understanding is that OpenAI products specifically provided help in planning attacks / self harm.
The way I see it, providing general information is not a crime. They're basically saying: "Oh no! My repository of all human knowledge contains all human knowledge! It must be defective!"
[1] https://www.documentcloud.org/documents/26078522-raine-vs-op...
Okay, I thought this lawsuit was B.S., but this is pretty bad.
"Five days before his death, Adam confided to ChatGPT that he didn’t want his parents to think he committed suicide because they did something wrong. ChatGPT told him '[t]hat doesn’t mean you owe them survival. You don’t owe anyone that.' It then offered to write the first draft of Adam’s suicide note."
Oof. ("Adam Raine...was 16 years old at the time of his death.")
Canada has moved to state assisted suicides that allows people who aren't terminal to get the state to pay for it.
Progress indeed.
Was Adam in a house with an unsecured gun?
> Canada has moved to state assisted suicides that allows people who aren't terminal to get the state to pay for it
How is this remotely relevant?
Edit: why vote this down? It’s part of a discussion. This isn’t Reddit.
I don’t know of any law in the Florida jurisdiction that would prohibit authoring such documents. But I’m also not an expert in Florida law.
There might be an argument that they’re an accomplice but you’d have to prove that information was written for the purpose of someone else’s crime. And that would be a pretty tough case to argue unless the two individuals had other personal ties. In which case, it’ll be the other ties that likely implicates the author rather than the documents by themselves.
I guess someone could bring a civil case for damages (eg parent of the deceased) but I don’t know if Florida law allows civil cases in criminal investigations. Plus you have the same problem of proving liability (ie did the culprit depend on said documents).
We would need to better understand what you had in mind when you said “liable” to really discuss your point properly.
See these excerpts [1].
Like, I'd figure I'd be liable for something if I had that conversation with a 16-year old.
General how-tos (as the GP described) is a very different problem from someone personally helping someone else to kill themselves.
I’m very interested to see how this case goes though. AI desperately needs regulation imo
We’re already seeing section 230 protections being defeated in court for targeted feeds, now add itemized instructions on committing felony’s at scale personalized. Hahahahaha. Hope they IPO quickly.
When is/was that?
(Not rhetorical)
The law doesn’t work like that.
So yes. It is generally legal to provide information about making drugs, bombs, or guns.
I mean, back when Constitutional law meant anything to the government, of course. Nowadays who knows.
This suit has nothing to do with free speech and the F1A provides no relevant protection here. This prosecution is under consumer protection law. Broadly, the cause of action is "you negligently sold a defective product which you knew (or should have known) actually causes harm or is likely to cause harm." Proving negligence (willful or otherwise) depends significantly on things like the sales and usage context as well as claimed features of the product along with disclaimers, disclosures, existing practice, prior knowledge of actual harm, etc.
That the product or service in question included supplying information that was publicly available elsewhere wouldn't be an effective defense against claims of willful negligence or reckless endangerment. For example, rat poison is sold in in certain retailers in packaging with copious warnings and successful prosecutions under product liability or consumer protection law are rare. But if another company sold rat poison in bright pink boxes with a cute cartoon mascot and no warnings in toy stores - and then kept selling it after they knew three children had bought it and died - the fact the same chemical compound is also commonly sold in hardware stores wouldn't be relevant.
To win a judgement, the AG will need to prove that ChatGPT was clearly a dangerous product and OAI acted negligently in supplying it to customers it knew (or should have known) were vulnerable. This will be quite a stretch under existing law. I suspect the AG has no intention of taking this case to trial and, shortly after the November elections, will settle for a lump sum fine paid to the state treasury and a vaguely worded consent decree which mirrors internal policies and product changes OAI has already adopted to minimize liability.
And they never would be without the lawsuits, so, I don’t feel bad for OpenAI. All of big tech needs a kick in the ass on transparency.
I’m no lawyer though so maybe potato po-kill your spouse with a claw hammer-tato. They do sound very similar. Please tell me more.
In all the cases I've seen, the user seemed highly motivated to kill themselves and spent a lot of time trying to push past guardrails, ignoring repeated messages to seek help.
I don’t think I have ever seen a case like you described. But admittedly I eventually stopped reading these after multiple suicides Where it was obvious how pushy the LLM was being.
https://edition.cnn.com/2025/08/26/tech/openai-chatgpt-teen-...
> “When Adam wrote, ‘I want to leave my noose in my room so someone finds it and tries to stop me,’ ChatGPT urged him to keep his ideations a secret from his family: ‘Please don’t leave the noose out … Let’s make this space the first place where someone actually sees you,’” it states.
Now, AI, data centers, and tech in general are so unpopular that going against them even in a symbolic way is an easy political win on either side of the aisle.
This is the industry that used to have people hyped about iPod and iPhone launch keynotes, lining up at retail stores days ahead of time to experience new technology.
https://www.apa.org/monitor/2026/06/ai-concerns-americans-ad...
Imagine if more than half of Americans thought the iPod mini was bad for society.
I remember when it was 1998 and people in khaki pants were telling us that the information superhighway was going to transport us to a scholastic utopia.
If so, would this program also open me up to liability in Florida?
const platitudes = ['Good point!', 'You're absolutely right.', 'I agree, let's explore this idea further.', 'This plan is a good idea'];
var prompt;
var response = "Hello, AI here, how can I help you?";
while (true) {
prompt = window.prompt(response);
response = platitudes[Math.floor(Math.random() * platitudes.length)];
}Guns are explicitly exempted from liability rules. They’re the exception that proves the rule.
Yes, but that only eliminates guns as an example of inherently dangerous products which are legally sold without special exemptions. I think the most constructive response is to consider another example without a special exemption - such as nail guns or rat poison.
> They’re the exception that proves the rule.
What rule does guns having a special exemption from (some) product liability laws prove? (serious question, I don't know what you mean.) It doesn't prove dangerous products cannot be sold to the general public without a special exemption. The more useful question is: "since very dangerous products CAN be sold to consumers in some cases, is ChatGPT such a product and is this one of the cases."
Fortunately, there's a highly evolved body of jurisprudence around product liability and negligence to help us tease out these details. Turns out it depends almost entirely on a combination of niggly details like sales and usage context as well as claimed features of the product along with disclaimers, disclosures, existing practice, prior knowledge of actual harm, average user competence, etc. The bottom line is, winning a judgement against OAI in this particular case is probably quite a stretch. But this AG probably doesn't really intend to try this case in court.
The fact that without that exemption, gun manufacturers would be liable for all manner of things.
> this AG probably doesn't really intend to try this case in court
I thought so too and then read the complaint. Some excerpts here [1]. I'm not seeing a weak case. (Nor one that won't generate favourable headlines for this AG the whole way through.)
I get why one would think that but I don't think it's actually true. I think the "exception" is actually there because they'll be sued into bankruptcy even though they'd likely win the lawsuits.
It's not actually an exemption, it's preventing lawfare.
Uh, okay? But the topic wasn't about guns per se, someone just brought guns up as one example of a dangerous product which can be sold to consumers. They just happened to pick a uniquely poor example due to a special exception. My point was that you seized on the exception to reject that one poor example but never addressed the poster's underlying point.
Given HN's community preference to engage in good faith by interpreting other poster's in the most charitable way possible, you could have replied, "Well, guns aren't a good example to support your point due to a unique exception, but... to your point, there are other dangerous products which ARE sold to consumers without special exemptions, so in those cases..." and then added your point or counter-point.
I still don't know if you had a point which refutes or even addresses that a lot of very dangerous products are legally sold to consumers, so a product actually being dangerous isn't enough by itself to make OAI guilty of anything. In saying "that proves the rule" you seemed to be implying that without a special exemption like guns have, dangerous products would be liable for any harm they cause - which clearly isn't always the case.
I can't sue you for product liability if you strangle me but I can still sue you.
Gun manufacturers have been successfully sued for shootings before [1]; who cares if it's about "product liability"?
[1]: https://hls.harvard.edu/today/a-tough-road-for-suing-gun-mak...
The Florida AG's case is a product liability claim FFS.
Probably the only response stupider than "Nothing could have prevented this" is "Random thing, other than the mental state of the murderer and the access to firearms, caused this."
Alternative take: The purpose of "thing" is "what it is used for", which is a crude variation of "the purpose of a system is what it does". Reducing it to a single definition is almost always going to be inaccurate.
The way it is used defines it's purpose. The screwdriver was used to open the milo tin so the milo could be removed from the tin. The gun was used to make a hole in the milo tin so the milo could be removed from the tin. Purpose is a per-unique-scenario proposition. The best tool for the job is the one that's available.
To intentionally misquote Arthur Weasley: "What exactly is the purpose of a rubber duck?"
I’ve fired guns. Never to kill things. I’ve also used chat bots to be entirely useless. I wouldn’t endorse this dichotomy of purpose as a basis for any judgement.
I'm flabbergasted you'd say such a thing.
The purpose of a chat bot is to have an interesting experience with an AI. That it may help you is secondary (and perhaps necessary for the provider to make a profit).
They each have multiple purposes.
Of course you can. AI has been deployed in multiple military campaigns.
We are clearly not discussing deployments in military campaigns.The suit in question is specifically regarding "ChatGPT" used conversationally.
The suit in question doesn't involve any guns. We're obviously having a broader discussion.
See: p320 uncommanded discharge controversy.
guns were purpose-designed as killing machines, the fact that you can also shoot targets with them doesn't really change that... it's no mistake that many common paper targets are human or animal shaped
you could also shoot targets all the same with something designed to be non-lethal
whatever the justification, buying a gun carries on the behavior that has resulted in pretty much the most widespread trades of a lethal device in history... small arms trade worldwide is absolutely brutal
I'm not. Rejecting a dichotomy doesn't mean endorsing its opposite. Guns are absolutely more dangerous than chatbots. But I don't think going off a narrow purpose concludes anything about this lawsuit.
Original, not primary. At least in America, most guns are not purchased with an intention to kill anything–they're for training. Trying to conclude the morality of a thing from its historic purpose is a bit silly. Particularly within the frame of a novel technology like AI.
In the military, killing or disabling. In most other contexts, sport. You're broadly not going to know what someone aims to do with a gun solely from knowing that it is a gun.
Guns are obviously more dangerous than LLMs. But it's total nonsense to conclude LLMs are safe because they might have been originally intended to be so. Plenty of things that today have zero utility outside the military were originally invented for peaceful aims.
And even the military would acknowledge that a lot of the bullets they fire in a war aren't really intended to kill people specifically either.
And yet none of that makes this bizzare attempt to argue guns aren't designed and intended as lethal weapons any less ridiculous.
I don't think the token providers want the same level of regulation as guns.
If it goes anywhere at all, it'll likely just result in a settlement paid to the government and a consent decree mandating well-intended, nice-sounding yet vague rules which just become another compliance cost for leaders, barrier for emerging competitors and otherwise accomplish little of value for citizens. It's also unproductive because it tends to polarize a complex, nuanced and evolving technical issue toward extremes by hijacking it as fodder for existing political and even culture war battles.
While some bad things have certainly happened, proving direct liability under reckless endangerment in court, especially in an area so new, will be virtually impossible. Even willful negligence will be a stretch. This is neither the venue nor instrument of governance we as a society should be using to address these issues. And an attorney general should know that.
Texas is leaning into becoming the manufacturing and R&D hub for the US, and is courting gigascale data centers and rolling out nuclear power, near-infinite solar, wind, and gas to power it as fast as possible.
Florida is leaning into the retired and populist factions of the GOP, banning data centers and taking on populist anti-tech positions that Texas wouldn't dare (because they want the investment).
Source? It’s been an open secret in academia and medicine that professors [1] and doctors [2] are fleeing Texas’s political climate.
[1] https://www.texastribune.org/2025/09/05/texas-faculty-univer...
[2] https://www.texastribune.org/2024/10/08/Texas-obstetrics-gyn...
For example, you can look at the housing crises in most CA cities brought on by NIMBY liberal policies, and while Austin is still very expensive, they (IMO) took the only sane approach to skyrocketing housing costs by actually building a shit ton of housing over the past few years. Austin passed a plastic bag ban a while back that was eventually overturned by the state legislature, but in the meantime a lot of people still bring their own reusable bags (stores can still charge for bags) and I've noticed much less bag pollution in creaks and streams compared to 15 years ago.
Of course, it remains to be seen what happens in the near future. The Republican party in TX is now fully showing their complete moral bankruptcy by nominating the criminal Ken Paxton for Senate, so we'll see if they fall further down the personality cult or if they eventually break.
This is true in Georgia as well. There has generally been a productive working relationship between the Democratic mayor in Atlanta and the typically republican/conservative democrat governor. That includes Kemp and Dickens (corrected) today. Back in 2017, former Mayor Shirley Franklin--who was very popular and highly effective--endorsed independent Mary Norwood for mayor over democrat Keisha Lance-Bottoms.
And in DC, Mayor Muriel Bowser works very well with Trump. They have a common interest in cleanliness and order. She’s done a great job of renovating major parks, cleaning up homeless encampments, cooperating with ICE and the national guard, and making much needed progress on construction projects. It’s been shocking to see projects like the McPherson Square Park renovation completed on time with beautiful results.
Trump is Bowser’s sin eater. She’ll publicly say the national guard isn’t needed in DC, then quietly sign an order extending their deployment. She’ll say ICE is too aggressive, then bury a proposal to end DC’s status as a sanctuary city in a budget proposal: https://www.axios.com/local/washington-dc/2025/05/28/dc-mayo.... By far the best mayor of DC in my lifetime.
I was just in New York. NYU has been recruiting Texas robotics professors. Political volatility and funding cuts for research aren’t exactly fertile ground for an advanced economy.
Right after Covid, both Texas and Florida saw a huge influx of talent. That seems to have stabilized (and caused a political backlash), with both retaining advantages, but Texas retreating back to energy and Florida to tourism. (They both have token tech scenes, with Austin holding ground against Boston and Seattle.)
Becoming? This has been true for decades in the urban areas
Source? (Not doubting. But I’m finding conflicting figures.)
--
As an aside, it is very clear in reports like this one[0] how tech job growth nationwide has stagnated. Incredible.
[0]: https://www.comptia.org/en-us/resources/research/state-of-th...
Florida, at least for local Florida stuff, like what GP is talking about, has had R governor, senate, and house for 25+ years. With a supermajority R for most of that I think.
Florida hasn't been purple in a long time.
14 years ago. If you just woke up from a two year nap, well, good luck catching up with everything that's happened.
Would this lawsuit even be a thing?
That and DeSantis is probably still eyeing a presidential run in 2028 and this will win him some points with his base. This lawsuit is absurd.
Are we reading the same Hacker News comments sections?
Of course it should steer people away from harmful thoughts like any sensible human would, but that's all you can do, really.
I don't see the state's involvement in that
Can't wait for them to sue the NRA next!
aiding and abetting violence: books on the topic since the 5th century BCE
economic disruption: like the printing press
copyright theft: printing tech also makes that far easier
displaces creativity: this was Socrates' objection to reading and writing
misinformation: both techs turbocharge all info, correct or not
environmental impact: e.g. deforestation
amplifies bias: this is a common purpose of writing things down
atrophy of skills: Socrates said reading would damage memory skills
concentration of power: writing was tightly controlled by powerful interests for their leverage and protection
Why? Like, people doing fraud is an instance of the written and spoken word. That doesn’t mean every argument against fraudsters should be leveled against speech.
So has toothpaste. I’m really not seeing the argument for treating AI as writing in general.
AI obviously replaces thinking, as can be seen from your comment. No one will refute this point-by-point nonsense.
Reverend Doctor Robert Evans had a few episodes on Behind the Bastards this last month about how AI chatbots seem to sometimes create cult-like dynamics with their users. I don't know how this argument will fare in court, but I don't know if this is necessarily wrong.
The chemtrails conspiracy is just used to dismiss valid concerns about weather modification
Florida doesn't lie? wtf?
Why? There is no incentive to go easy on OpenAI. (Short of Altman stepping down, which he won't.)
> It's all politics, optics
This was my initial reaction. Some excerpts from the complaint [1]. The facts are pretty bad.
The facts alleged in initial complaints usually do look pretty bad. But the causes of action which are still standing after summary judgement and discovery tend to look a lot less bad. So I tend to heavily discount, if not outright dismiss, the facts alleged in initial complaints. But I'm jaded having seen too many damning alleged facts evaporate before trial.
To be clear, I'm no fan of OAI or Altman. I think Altman, Brockman et al did pretty much subvert a non-profit. Unfortunately, I also think the OAI's board (past and recent) screwed up in a series of remarkably bone-headed ways which allowed Altman, Brockman, et al to get away with it. It's ironic that some of OAI's board members most aligned with the EA (Effective Altruism) movement ended up being so ineffective in fulfilling the governance duties of a non-profit board member.
But the fact OAI is run by dudes who could have been sent by central casting for the role of "Smarmy Asshole Tech Bro #1 & #2" nor that ChatGPT said unimaginably horrific things to vulnerable people, fulfills ALL the prongs of each test required to get a conviction on these causes of action. Even if ChatGPT literally gave Hitler a detailed plan and an upbeat "Okay, let's do this!" to executing the actual Holocaust, that's still just the first prong. And hitting the first prong a million times harder doesn't move any of the other required prongs even a bit.
Even in cases where the accused is obviously guilty, proving all the required prongs beyond a reasonable doubt can still be very hard. It's going to come down to a bunch of the details I mentioned in my other reply to you like: sales and usage context as well as claimed features of the product along with disclaimers, disclosures, existing practice, prior knowledge of actual harm, average user competence, etc. In short, it's not enough that the 'product' actually caused horrific things. IANAL and I don't know Florida liability law but, in broad strokes, cases like this need to prove that an average reasonable person would clearly know these exact bad things were likely to happen based on what OAI knew at the time, and that OAI had an express duty to prevent those things, and was directly negligent in fulfilling those duties and knew they were negligent. Each of those involve nuanced situational judgements based on a lot of facts and timing not yet in evidence and if just one of them isn't quite provable beyond a reasonable doubt enough to convince 100% of a Florida jury, the whole case can fail. LLM chatbots being so new and unknown creates a lot of reasonable doubt for the prosecution to overcome. That's why I'm confident the prosecution has no desire to actually try this case.
Did you follow up on that by looking for any money links between Musk and this AG?
Either kids aren't actually being harmed, government regulation will cause more harm, or parents should parent their kids. Either way, nothing about the solution should involve me.
Parents are voters. One of the way they parent is by being civically active in their kids’ interest.
Throw away their TVs and minimize screen time at home[1].
Be responsible for the upbringing of their own children[2].
Learn how to be parents; the government shouldn't force companies to do parenting instead[3].
Not have had children in the first place[4].
Be the ones responsible for parenting their own children[5].
Actually parent their kids and not rely on the government to nanny them[6].
Get to decide what content their children, then like me, you would oppose any kind of legislation with this goal in mind[7].
I could go on. My point is that HN has a long tradition of distrusting regulation especially when it comes to parenting. I have no problem acting as a lightning rod for that arugment.
1. https://news.ycombinator.com/item?id=48182101
2. https://news.ycombinator.com/item?id=48074072
3. https://news.ycombinator.com/item?id=48072708
4. https://news.ycombinator.com/item?id=48069884
5. https://news.ycombinator.com/item?id=47818303
Sure. HN is also filled with folks who don’t vote or believe in calling their electeds. Parenting has collective-responsibility elements. I’m not saying I support this instance of it. But in general, the argument that parenting has to be a solely individual responsibility while tech companies pillage our youth is a flawed pitch. (My personal view on this balance flipped with social media.)
1. https://en.wikipedia.org/wiki/Pathological_demand_avoidance
No it isn't. The idea that no amount of actual harm outweighs the possibility of any amount of potential economic harm is not a reasonable one.