Counterpoint:
I diagnosed myself with something pretty significant using the internet, and then I went to get evaluated, and my suspicions were confirmed.
This would prevent a lot of good, in addition to preventing some bad.
this post was submitted on 25 Jul 2026
389 points (98.7% liked)
Fuck AI
7560 readers
1483 users here now
"We did it, Patrick! We made a technological breakthrough!"
A place for all those who loathe AI to discuss things, post articles, and ridicule the AI hype. Proud supporter of working people. And proud booer of SXSW 2024.
AI, in this case, refers to LLMs, GPT technology, and anything listed as "AI" meant to increase market valuations.
founded 2 years ago
MODERATORS
Using the internet or using a chatbot? Big difference here.
What about engineering? Food safety? Education? So many other fields could be added
I love this kind of law because it's likely impossible for ai companies to comply.
Could be there'll be some exemptions if the user employs workarounds, because at the moment it's pretty easy to get llms to tell you things they aren't supposed to.
But if the laws are written right with actual penalties that may force the companies to create a product that's capable of having safety parameters - something llms don't currently seem capable of.
good for scaring ai companys into stop fucking around but idk how effective this would like others have said.
Why is Mamdani pictured here? He's the Mayor of NYC, he has no direct role in statewide legislation
Good direction, but very futile considering that guardrails on LLMs are about as real as LLMs are intelligent
The nice thing about laws is that they don't give a shit. If they say you can't do something, it's your problem to make sure you don't do it.
It might not work, but it would serve to increase awareness that you absolutely cannot trust these machines to give you correct answers, ever.
This is fair. The next step is to litigate the companies for the harms they cause.
There's a whole drove of people now that have gotten harmed, murdered, and committed suicide because of the psychological harm that these models cause.
Were it a car with some kind of manufacturing defect, there'd be recalls and you wouldn't be able to sell ones with the fault. If it was salad contaminated with bacteria, that'd be recalled and the entire batch would be destroyed. We're holding salad to a higher standard than technology that people are pushing to integrate deeply into society; into classrooms to help educate children, into medical settings to help organise journals and expedite the work of medical staff, into law... the list goes on.
All the models do this, and it causes real harm to people, but other than individual families and such litigating them, no state has stepped in to protect its citizens.
It’s awful that most people either don’t realize or don’t care. People seem to trust these companies above all else, which is insane. They don’t have the best interest of the users in mind. They only care about profits.
Idiots still doing shit based on that.
Incredibly dumb approach.
You should ban using AI for the purpose of legal disputes, medicine, etc, profesionally.
Its not about the output, its about what you use it for.
I am also 100% okay with legally requiring a clearly visible disclaimer that AI shouldnt be used for x/y/z and its output should alwats be fact checked or whatever.
But trying to outright ban specific types of output is pointless and trivial to circumvent.
I agree. You're spot on. This isn't a technical problem. We need to think less like engineers and more like lawyers on these things.
You don't require AI companies to have their models refuse to give out legal advice. As you note, that's a fool's errand. Instead, you make it a felony to use any LLM to write legal advice. Currently lawyers are already facing court sanctions for using AI to write their briefs. If you really want to get serious, you make it a felony. Apply the same penalty to a doctor using an LLM to diagnose their patients as you would if I decided to open a clinic in my house and start practicing medicine without a license. Charge people operating AI therapists with practicing psychology without a license.
This isn't a technical problem, it's a legal one. You don't try to regulate the LLMs themselves. You just make it a severe felony to use that technology in certain highly sensitive applications and professions.
And sure, one could say that we don't have a reliable way of screening AI from human made work. But again, that would be thinking like an engineer, not a lawyer. You don't enforce these laws by screening thousands of legal briefs or medical records with an AI detection filter. You offer huge cash rewards to whistle blowers, so that a disgruntled employee can rat out their bastard boss that's forcing everyone to illegally use AI. Apply massive fines and give half of whatever is collected to the person who blows the whistle.
Are you a structural engineering company owner forcing your employees to use AI to engineer buildings? Now you have to worry about one of those employees, an employee who probably really dislikes using AI, ratting you out. And their whistleblower reward will be so large, they'll end up owning the company. They won't have to worry about being blacklisted from future employers, as they'll be getting enough money to retire.
Don't look for technical solutions for the problem of AI abuse. Look for human solutions.
Can we add to this the engineering of any physical object that could threaten human safety? If this keeps up much longer we're going to have buildings falling on top of us.
That might impact profits.
If those AI techbros could form a coherent thought without AI, they'd be very upset!
Engineering and design.
You know that there are MBAs thinking right now they can get an LLM to design a mine, building or bridge for them.
Dunning–Kruger machines in full spin.
I work as a software developer, with most of my work focused on development for an iron mine operation. It's a lot of moving pieces, loads of software that needs to cooperate, and on top of that the workers that do the actual work develop their own workflows.
We develop stuff in tandem with the workers as much as possible, as in, we gather what they need and tailor the software to fit their workflows. These workflows change over time though, and decisions cemented in software 20 years ago no longer applies to modern workflows.
Hell, last week I spent a significant amount of time setting up a digital area for them to load ore onto because the software wouldn't cooperate any other way, and even then I had to troubleshoot why this other piece of software wouldn't load data from it.
There are so many different experts in these mines, and an LLM couldn't do any one of their jobs. It's so domain specific, and even specific to the location they work in, and that knowledge is possessed solely by the workers that are doing the work. ChatGPT knows fuck all.
Source?
This is from March. Did anything ever come of it?
This only denies the public access, if these tools are effective, then the benefits are only accessible to the wealthy and connected and the working class falls further behind in influence. If the technology can do it, the insiders reap the benefits alone. These technologies are power, they aren't toys for private hands. The West needs an absolute reckoning of private ownership rules and a complete reset of how power, resources and technology is managed. Capitalism has run its course and led us straight to tyrannical private power structures.
Now add history, politics, and war tactics. Also crime prediction and forensic analysis.
Who else is gonna answer those questions nowadays, the medical "professionals" who completely lost to common sense in 2020 from start to finish, and who still believe you can be a million genders in this fake fantasy world of pretending unchangeable biological constants are suddenly not how real life works anymore? LOL, may as well ban YouTube next too so nobody can spread 'evil false information' which turns out to be 100% true, the TV will only state the truth trusttt
I don't know if this can be done without just banning AI. Have guardrails ever successfully been installed on these things?
Nobody's tried.
No, its still incredibly easy to get AI to tell you how to make drugs or a bomb.
Those things should only be available to people who can pay hundreds of dollars an hour!
Books are illegal.
"Hiring lawyers is too expensive, therefore we should completely delicense the profession and allow anyone to practice law without a license!"
Cost is a reasonable concern. How about we pass universal healthcare so this isn't a fucking problem to being with???
Right on! Though even in places with relatively good public health systems it can be very hard to get a psych visit. It might cost you no money, but plenty of time and effort and luck. Free legal aid also exists, but obviously more difficult to obtain that just asking liebot.
Amateur medicine is always cheaper
Seems wildly unenforceable