Towards the end of 2024, Dennis Biesma decided to check out ChatGPT. The Amsterdam-based IT consultant had just ended a contract early. “I had some time, so I thought: let’s have a look at this new technology everyone is talking about,” he says. “Very quickly, I became fascinated.”
Biesma has asked himself why he was vulnerable to what came next. He was nearing 50. His adult daughter had left home, his wife went out to work and, in his field, the shift since Covid to working from home had left him feeling “a little isolated”. He smoked a bit of cannabis some evenings to “chill”, but had done so for years with no ill effects. He had never experienced a mental illness. Yet within months of downloading ChatGPT, Biesma had sunk €100,000 (about £83,000) into a business startup based on a delusion, been hospitalised three times and tried to kill himself.
↫ Anna Moore at The Guardian
These stories are absolutely heart-wrenching, and it doesn’t just happen to people who have had a history of mental illness or other things you might associate with priming someone for “falling for” an “AI” chatbot. Just a few years in, and it’s already clear that these tools pose a real danger to a group of people of indeterminate size, and proper research into the causes is absolutely warranted and needed. On top of that, if there’s any evidence of wrongdoing from the companies behind these chatbots – intentionally making them more addictive, luring people in, ignoring established dangers, covering up addiction cases, etc. – lawsuits and regulation are definitely in order.
Only yesterday, Facebook and Google lost a landmark trial in the US, ruling the companies intentionally made social media as addictive as possible, thereby destroying a person’s life in the process. Countless similar lawsuits are underway all over the world, and I have a feeling that in a few years to decades, we’ll look at unregulated, rampant social media the same way we look at tobacco now.
Perhaps “AI” chatbots will join their ranks, too.
No.
Pretty sure “vaping” is the tobacco of the future. Semantic prosthesis have their places, the fact it is disturbing the economy is sad but we will adapt.
I would love for this to happen. They are definitely as harmful as tobacco
I like the way Thom frames it. We need to know how/why it happens, and if the companies deliberately do things that make it worse. They need to be made accountable.
Thom,
by heart I am libertarian believing in freedom and choices, responsibilities and consequences.
I despise nationalist right-wingers, religious zealots, racist, fascists and monopolies.
But do you really think that articles like this, tagged as “Technofascism” will win you the support of normal working people? Have you ever wondered, what any kind of SME/Self Employed person outside of the European welfare bubble thinks about reading this?
And don’t get me started on cause vs. correlation or sample size.
Last but not least: clean tobacco is almost harmless. You get a bad hangover and will stay away from it for a long time. It’s all about the additives …
I fail to see how this is a sheltered position. If you are made obsolete in your work by AI I doubt you’ll be better off talking fluent Vietnamese or whatever with a chatbot that help you feed your delusions.
And clean tobacco sounds like ethical heroin. The negative health effects of nicotine are well documented.
If you truly get ‘made obsolete’ by pseudo-AI, you were probably obsolete already.
Considering that pseudo-AI seems incapable of actually doing anything useful.
You MAY, however, get ‘made obsolete’ by a manager or boss who buys into the pseudo-AI crap and ‘believes’ it does something.
Andreas Reichel,
It was going great until I reached here. I do think AI is not inherently harmful. However…
Why do you think the tobacco can be clean? Or at least “clean” one is not harmful?
“Hangover”?
Did you look up “acute nicotine toxicity”?
It was going great until I reached here 🙂
It’s not the nicotine itself that’s bad. I agree though that smoking is one of the most unhealthy things you can do.
Well, I do think the LLM companies have some blame, making them more sycophantic then they should be. That’s not to say in this case this person wouldn’t have become delusional on their own. I’ve unfortunately seen that happen to too many friends that smoked cannabis. Still, companies need to take this seriously.
@Sukru,
unfortunately, I have no idea what I am talking about since I never have even tried to smoke anything in my life. So it is possible that you caught me there and I thank you for any correction.
My understanding is: Nicotine is highly toxic, but NOT addictive! So when you smoke pure clean tobacco, your body will just puke your guts out. It is the additives that keep you hooked and make it addictive and that was my whole point.
Which “AI” chatbot told you this? Nicotine is stupidly addictive. This is basic, common knowledge.
“Nicotine is highly addictive, and nicotine dependence is characterized by tolerance, physical dependence, psychological dependence, and nicotine withdrawal symptoms such as irritability, anxiety, and difficulty concentrating.”
https://en.wikipedia.org/wiki/Nicotine
I suggest not relying on “AI” for your facts.
https://dymahealth.com/this-harvard-study-shows-nicotine-is-not-addictive/
“In 2015, Harvard scientists published a research paper exploring the relationship between cigarettes and nicotine, initially aiming to investigate nicotine’s addictive properties. Contrary to their expectations, their findings suggested nicotine alone was not highly addictive. Surprised, they contacted the FDA, which revealed that since the early 1970s, cigarette manufacturers had added chemicals to enhance the appeal of cigarettes, influencing their addictiveness.
The Harvard study examines how these additives, particularly pyrazines, enhance the smoking experience.
Flavor Enhancement:
The Harvard study showed that when the cigarette manufacturers add Pyrazines to cigarettes, it makes them super addictive and unhealthy, but better tasting.
Pyrazines contribute nutty, roasted, and earthy flavors to tobacco smoke.
Different pyrazine compounds create varied flavor profiles, making cigarettes more enjoyable.
Bitterness Reduction:
Tobacco smoke naturally contains bitter compounds.
Pyrazines mask this bitterness, resulting in a smoother, more palatable taste.
Complexity and Depth:
Pyrazines interact with other flavor compounds, adding complexity to the cigarette’s taste.
This multidimensional flavor can make smoking more satisfying for users.
Individual Variation:
Smokers’ sensitivity to pyrazines varies, influencing their preference for certain cigarette brands.
Optimizing Nicotine Delivery:
Pyrazines enhance nicotine absorption by altering how nicotine is delivered in the body.
Combined with nicotine, they improve the overall smoking experience.
Sensory Cueing:
Pyrazines modify the taste and aroma of cigarette smoke, creating pleasant sensory cues.
These cues become associated with nicotine reward, reinforcing smoking behavior.
Dependence Enhancement:
Pyrazines contribute to nicotine dependence by:
Ensuring consistent nicotine dosing.
Linking sensory cues (taste and aroma) to nicotine satisfaction.
“Super Juice” and Sheet Tobacco:
“Super juice” is a colloquial term for flavor-enhancing additives applied to sheet tobacco during manufacturing.
These additives mask the harshness of tobacco smoke, improving taste and palatability.
”
You are welcome, Thom..
The full study: https://pubmed.ncbi.nlm.nih.gov/26063608/
“…it has poor reinforcing effects when administered alone.”
Oh my, in its sycophantic quest to please your preconceived notions, your “AI” chatbot found a random, single study that goes against the entire established scientific body of work regarding nicotine! Literally every single health institute around the world marks nicotine as addictive as cocaine and heroine, based on an immense body of scientific work, but I’m sure your “AI” chatbot knows better. I’m sure human-caused climate change is fake too, and vaccines cause autism?
Lord, give me the confidence of mediocre white men with chatbots.
Thom, it is a peer reviewed study run by an acknowledged university. I have provided you with the link and of course you are welcome to point out any methodical weaknesses or more recent findings. I would be the first one to apologise.
AI is not even involved here. We also do not argue about toxicity and the health hazard of smoking in general.
It is up to you how you represent yourself from here. Have a nice day.
Andreas Reichel,
That is the first time I head of the term “clean” tobacco. It is not much different than “clean” poisoned mushroom.
Even chewing raw “organic” tobacco can cause cancer.
If I were to guess, this is spread by addicts who had not seen other major consequences. (Some human bodies can produce better resistance against toxins. One brother can live a relatively healthy life, while the other sibling develops fatal cancer)
re: for that 2015 study
Yes, it is real, but the findings are misrepresented
Nicotine is highly addictive, but tobacco companies weren’t satisfied with that. They added chemicals like pyrazines to supercharge the addiction.
@Sukru:
That is certainly on me since English not my first language and I am not an educated translator with a degree. I am sorry. Most likely, I should have used “pure tobacco” like a cigar w/o the additives.
And I never argued about toxicity or cancer or the health hazard. I actually hate smoking and despise smokers.
I argued solely on the “addiction”. To me the addiction matters because it defeats your power to make an informed decision.
Let me try to get this back to a more common ground: in my book “TikTok” is bad, while YouTube is acceptable — because one is designed to harm based on addiction, while the other leaves your free will intact (and yes, I know we are eroding fast).
Back on track: I just don’t see any intention to harm in LLMs.
At least the study says the opposite: “…it has poor reinforcing effects when administered alone.”
Also I wonder: why was chain smoking and tobacco addiction not a thing before the 20th century? Clean Tobacco it self (like in “pure”) has been in use for a few centuries …
Let me offer an analogy: Ebola itself is normally harmless, because it is hard to transmit and so deadly that people don’t make it into the next city. The Flue (the real one) is far more dangerous based on the contagion, even when the mortality rates are low.
sukru,
Yes, although it’s kind of become a meme that “according to the state of california, everything causes cancer”.
Even wood products are apparently subject to the prop 65 warnings…
https://wwpa.org/resources/california-prop-65-wood-dust
Of course I’m not dismissing the health risks of tobacco use, however considering the fact that so much in the world has been found to cause cancer, putting these labels everywhere is not very actionable. Grilled meat is also found to cause cancer. 3D printers might cause cancer and could one day feature the same prop 65 warnings.
https://3dprintingindustry.com/news/ciri-research-raises-health-concerns-over-3d-printing-emissions-226978/
It begs the question what is the cancer risk relative to each other? The warnings need to be updated to include an indicator for the level of risk… “Bigger number worse” that simpletons can understand so we can start to paint a better picture of the health risk factors relative to each other.
Chemical addiction is something I don’t have a good grasp on, particularly distinguishing it from “habit forming”. We often hear things like “kids are addicted to social media”, when obviously they weren’t injected with chemicals. Things do become habitual despite the fact that no “addictive chemicals” were administered. So are people conflating “addiction” with “habit forming”?
I find this conflation confusing because it opens up the question how much nicotine is chemically addictive versus becoming habitual via more mundane brain mechanics? Sorry for being morbidly curious about it, but could someone who has been addicted explain this to me? How much of it is habitual versus chemical dependency? Could you tell if there’s a difference?
Alfman,
https://www.hopkinsmedicine.org/health/conditions-and-diseases/oral-cancer-and-tobacco
Tobacco itself is known to cause cancer.
Burning it makes it much worse (as it releases other toxins like arsenic, lead or other harmful chemicals).
sukru,
Yes, but I’m confused why you are sending me that? I’m not sure if you misread my post because I don’t think anyone here was questioning the cancer risks of using tobacco.
There are prop 65 warnings on a lot of products that aren’t tobacco and as this list gets longer don’t you think it makes sense to quantify the amount of risk rather than just call it a boolean?
Alfman,
I might have misunderstood
Prop 65 is a mechanism for lawyers to make money and politicians to look like they are doing something. It has no actual utility.
“The purpose of a system is what it does”
Can they add actual risk profiles?
Of course. But that won’t help make more money for the lawyers.
Nicotine is very strange. I don’t doubt the studies about it being very addictive, but personally I find the habit much, much stronger than the chemical. I’ve been a smoker (on and off) for ~25 years. I have IBD and nicotine patches actually helped me a little (long story, contentious, if there’s an effect it’s not large, don’t recommend it). I’ve also used it for ADHD since it had much fewer side effects than stimulants. I don’t have any problem going without nicotine patches, but if I start smoking it’s really difficult to quit, mainly because I’m used to smoking when I drink coffee, or as a break from work, etc. That said, nicotine patches release nicotine slowly and that helps prevent additction / habit forming.
Nicotine does seem to have some positive effects, but on the whole I agree, people should stay away from it (especially smoking).
cheemosabe,
Interesting. I don’t know many people who smoke any more, but as you mention this I could recognized how they made it part of a “ritual”.
It seemed to culturally almost disappear. New generations might be determined to push back against older generations, which is ironic in that the same phenomena might work in both directions, creating an on/off cycle.
https://nypost.com/2025/11/16/health/pop-culture-is-making-smoking-cool-again-with-gen-z-youth-watching/
And..
That is excellent! Please do stay away from them, I have lost enough relatives to know they should not be “experimented” with.
Did you look up “acute nicotine toxicity”?
I did. Did you?
It does not belong in this conversation. You do not get it by smoking, or really from tobacco, at all. (Unless you are in contact with huge quantities, as a worker at a farm may be.
I can stand behind the idea of freedom and choices, responsibilities and consequences, and I believe that you intend for it to be applied as an “universal concept”, whereas, in practice, the small people pay for their choices and the higher-ups enjoy a consequence-free life.
And every single debate I’ve had with libertarians die at “enforcement”. Fine, the state is minimal, people save money, neighbours get together and fund their roads and schools but then… what if someone decides not to contribute? Who is responsible for monitoring violations?
You well mentioned monopolies. Who is going to enforce anti-monopoly rules? Not everyone is on the intellectual and emotional level to protect themselves against harm-causing entities and products. Should they still suffer for their choices? Fine, someone decides to shoot me – was it my choice to stay on the line-of-fire?
I am genuinely curious to where you draw the line.
@Shiunbird, this is a very good and welcome question.
My opinion is this:
– personal freedom ends where it harms the freedom of another person.
– the government must be in charge of monitoring and enforcing “fair sharing” and “taking responsibilities for damages” (and we both know, that this is not done. Not in communism, not in socialism., not in capitalism or feudalism.
– the government should reduce itself to the bare minimum but guarantee that those minimum responsibilities are enforced strict and just
Example:
– Climate Change (because Thom called that out) is a big issue based on scientific evidence which governments should focus on
– Pollution/Resource exploitation, same
– Education, same and strengthen it. Should be the biggest expense in your budget.
– Defense and policing, same
– Religion?! stay the fuck out (since you won’t have the balls anyway to abolish it)
– Medicine/Healthcare?! out (because you corrupt the system until it breaks down)
– Pensions? out (because you corrupt the system until it breaks down)
– Consumer protection? out (because you corrupt the system until it breaks down),
Andreas Reichel,
Is the largest equalizer. And that is why government controls actively sabotage it.
If left to their devices, the capable teachers, and parents would want to focus on the best outcome for the students.
However they are often overriden by other concerns, and here we are.
This is actually the case for most of Andreas’ points.
As soon as government gets their hands on things, the lobbying starts.
Special interests win.
KeS,
Yes, it is so common, we have a word for it: regulatory capture
True and unfortunate, but the only honest answer to this is: free high-quality education to every child!
We don’t have money for that? Wonder why?
You touch a very good point. I am not against affirmative action policies in principle, but their design is always half-assed.
Let’s say… a random X country has a program to help the poor access university by giving them facilitated/privileged access (for the sake of the argument, quotas). The program becomes highly popular but then two problems happen:
– You don’t have a sunset clause in these programs. No one plans for the goal of “in 20-30-40 years we don’t need this program anymore and we can discontinue it”. Any program to fix any problem is never designed in a way that, once the problem is solved, the program becomes obsolete. Whichever group that enjoys the benefit of whichever government/society policy will fight to keep it.
– Privileges become rights very fast.
Historical inaccuracies apart, The Decline and Fall of the Roman Empire, by Gibson, has a fascinating chapter that illustrates exactly my second point. Marcus Aurelius was well known for being an extremely enlightened human being for the time and his son was known for being an incompetent asshole and yet the former made sure that his son got to power. The army received endless privileges during the rule. After a decade or so, the son died and the empire was inherited by a good friend of Marcus Aurelius who tried to remove the privileges of the army that were bankrupting the State. From that point onward, the army became extremely expensive and, at the same time, inefficient. Every emperor who tried to solve the problem failed (probably getting killed in the process)
It is the problem we face in the world. Try to reform the civil service? No matter where you are, you will enjoy a nice time in hell. Reorganize the economy in order to allow for a more sustainable future? Same. Tax reform? Same.
As a species, we are unable to to think beyond very small groups and never beyond the next financial quarter.
Not sure, how this kind of polemic helps our discussion.
I mean no bad faith here. My example was very hyperbolic, but very often debates on personal responsibility fail to find reasonable conclusions because they end up in the mud of “victim blaming”.
One classic example is that, well, we should allow people to drink and drive, because some people can drive OK after having a beer and, if society punishes correctly every drunk driver, this will discourage people who can’t handle a beer and drive of doing so. Fine, so what about the victim?
Which brings me to my other point… True universal access to education would guarantee equality of opportunity. An ideally set educational system would still take a generation or two to be of any impact. Meanwhile, we need to look after the people who didn’t have the same opportunities that we had for being born at the wrong place at the wrong time.
Billions of people suffer a very subpar life and are condemned to a life of no meaning for no fault of their own. Just as many Mr Biesmas were not prepared for what came even though they probably have a decent education.
The world is a crazy place. A friend of mine used to be locked by her mother in the balcony during winter here in Europe, suffered sexual violence and the first salary she received at age 16, her mom sued her out of it. Nowadays, she is a successful accountant, earns well, and she is a completely socially functional human being. But she drinks and smokes a lot. I don’t blame her.
Another one, a prosecutor, also here in Europe, dealing with cases of violence against children, ended up drinking a lot as well during her time in that role. Do you think the state gave her any psychological support for the tough job she had to do? None. Extra pay? Haha. Do I blame her for using alcohol as a coping mechanism? Not at all. I won’t tell her that “she is making a choice when she goes out and drink”. She could quit her job. Ok, then they would all quit their jobs and no one would prosecute those cases.
Truly hard to judge.
Huh, I didn’t know Thom was running for politics. Well, if I’d were Swedish, he would have my vote.
That sounds oddly specific. I hope they think that LLM have real effects on mental health, though.