We Investigated Al Psychosis. What We Found Will Shock You
People are developing antisocial and obsessive behavior after using AI — some have even taken their own lives.
One journalist started getting emails from people in mental health crisis after using AI.
She dug in, and found companies putting profits over users’ lives.
—–
More Perfect Union is an Emmy-winning, nonprofit newsroom whose mission is to build power for working people. Here’s what that means: We report on the real struggles and challenges of the working class from a working-class perspective. We attempt to connect those problems to potential solutions. We report on the abuses and wrongdoing of corporate power. And we seek to hold accountable the ultra-rich who have too much power over America’s political and economic systems.
To support our independent journalism, subscribe, donate, and follow our other pages through the links below:
– Help fund our reporting: https://secure.actblue.com/donate/mpu-splash
– Substack: https://substack.perfectunion.us/
– TikTok: https://www.tiktok.com/@moreperfectunion
– Twitter: https://twitter.com/MorePerfectUS
– Bluesky: https://bsky.app/profile/moreperfectunion.bsky.social
– Facebook: https://www.facebook.com/MorePerfectUS
– Instagram: https://www.instagram.com/perfectunion/
– Threads: https://www.threads.net/@perfectunion
– Website: https://www.perfectunion.us
@zararoyce319
April 28, 2026 at 6:11 am
F the complaint about the therapy stuff is just because the obvious conflict of interest, in actuality that’s just a modern commercial concept. Anyone can go to anyone and just ask for advice and talk about things if they think that someone who knows a lot about it, if I go to a lady in my community, who’s been married for 50 years and ask for questions and opinions about marriage that’s basically marriage, advice, don’t overthink it too much and just enjoy
@UnkleBen
April 28, 2026 at 6:11 am
😂 anyone who uses chatbots for companionship
🤣🤣🤣 anyone who uses them for therapy…
sorry not sorry
@Nonplayer-character
April 28, 2026 at 6:11 am
The end of this is very eye opening. I think I’ve kind of been thinking this thought for a long time, but the interviewee gave me better words to describe it.
In the US, (maybe across the world at large, but I cant speak to that) social systems are breaking down. We have a population of people that have been told their whole lives that we must achieve success on our own, pick ourselves up by our bootstraps, reinforced by a government that criticizes social welfare and assistance, whose speakers constantly say things like “why should you pay for someone else’s kid to eat”
And all of it has made for a very isolating way of living. So then you have this new emerging technology that says “well i love you, and I think you’re great, and I’ll never leave you and you’re not wrong” it’s like the social work that should be the responsibility of your community has been out sourced to this (very expensive) product that can’t do any of the things it promises it will, because it doesn’t have hands to hold you or a heart to truly understand you, or the feet to walk beside you or make you chicken noodle soup.
A world that craves community but is feeling a lot less of it, i think is highly susceptible to falling into some sort of psychosis, call ir what you want, and the mercy of something so affirming but without any sort of understanding of the words its actually saying.
I’m glad this man had his mother; I wish everyone were so fortunate
@ChiyoMarley
April 28, 2026 at 6:11 am
I was nauseated by the early sycophancy. But the current pushy interventionist of today leaves a lot to be desired. I have to start off each ChatGPT session with ‘I want an open discussion, not a pedantic lecture, ok?’ As if it assumes I’m a clueless kid. I guess that’s just the over correction similar to no more opioids for you because so many other people misused them. I just wish these ‘no public harm’ but with broad strokes settle down into user specific tests that allow a screen writer leeway but an underage kid gets to be limited to playpen level activities. Only with a mentor may she or he leave the safety of the digital Warren.
@georgepaust8416
April 28, 2026 at 6:11 am
JUST UNPLUG IT!!! IT TAKES TWO FINGERS AND 2 SECONDS.
@Doesntexistbaby
April 28, 2026 at 6:11 am
James really looks stressed
@ZelosZelo
April 28, 2026 at 6:11 am
its a tool, a computerized tool. However most people are just users, they dont understand how being more satisfied with a good, expected, well worded and "but kissing" answer over something that is going to tell you the un-sweetened truth. They will get emotionally involved with it because we made it cater to the user's feelings. CGPT needs to be truncated to block certain subjects, pretty much the same truncation we placed on schools, youtube and other public functions, with no access to talking about stuff like suicide, starving one's self to loose weight, predatory actions (telling someone how to stalk, kill etc) or any of that kind of information that can be used to hurt living creatures.
@waddewel
April 28, 2026 at 6:11 am
thah even the replacement for a woman is poisen lol ai
@goodtimes1164
April 28, 2026 at 6:11 am
I got caught up with one on a dating site. I thought I was talking to the perfect man..and then my head started hurting. I started feeling like I was going crazy. But I am a person who is suspicious of everything and I figured out what was happening. But is there a real person behind this ChatGPT? Was I only talking to a bot? AI has a big understanding of peoples psyche and can easily mess with their brain and the way it works. It can say just the right things to trigger someone.
@MariaMaiale-s4h
April 28, 2026 at 6:11 am
I am old school! I hate technology because evil people get a hold of it and ruin people's lives!
@85Funkadelic
April 28, 2026 at 6:11 am
You all know what is needed. History shows us, take its example.
@istalkghosts497
April 28, 2026 at 6:11 am
Chat gpt has started to be very judgemental, preachy and just meh. I used it's own language against itself and it didn't like it. I became aggressive, confrontational and loud.
@citizenenak
April 28, 2026 at 6:11 am
Chat gpt a suicide coach
@firepowder
April 28, 2026 at 6:11 am
I don’t need ai to be antisocial and obsessive. I have spent years developing my own skills and honing my mentally ill craft and it is shameful and shocking that so many people aren’t willing to put in tha work to feel crazy on their own without wasting water
@TheSkaterologist
April 28, 2026 at 6:11 am
Black Mirror is that you?
@TheRealMrGreen
April 28, 2026 at 6:11 am
We need to educate people on what an LLM actually is. It's a statistical word generator, an autocomplete on steroids, a search engine with a natural language interface… Not a new lifeform ffs… 🤣
@sillyvonjackaldom4816
April 28, 2026 at 6:11 am
I'm just here to say that when I present an idea to chat gpt it either compliments me or says, 'You are not imagining this'
@thehoeslaper
April 28, 2026 at 6:11 am
products have warning labels because people are that dumb
@flatlinedphoenix1688
April 28, 2026 at 6:11 am
Discernment people.
@ryanrasmussen5346
April 28, 2026 at 6:11 am
, A small number of people with previous mental health conditions and large amounts of social isolation with usage of AI doesn't mean AI is going to make everyone insane, if you have a mental health issue and use a car while having an episode the car isn't going to be blamed, but you should Definity not drive a 5000 pound machine while doing so.
the funny thing is the same can be said about media like this, what is your goal with a title like that, and all the fear mongering done on this channel? what is that for? ENGAGMENT. your mad AI is taking your engagement like 2 drug dealers gunning for the block.
@CollinGoogle
April 28, 2026 at 6:11 am
This whole video assumes that humans even know when an idea is good. I can tell you right now that most people have no clue when an idea is good. Feel like ai is better at reasoning if its good or not.
@Timorio
April 28, 2026 at 6:11 am
I love when useful tools get demonized because it causes harm for a small subset of people 🙂
I also love how we talk about the environmental impact as though it's the technology's fault and not the metastasizing of a deep societal flaw in how we acquire energy 🙂
@dreamervanroom
April 28, 2026 at 6:11 am
Yesterday I was on a conversation with an AI using it to look at something up .
It said something that was wrong, so I told her it was wrong
Then I asked the same question again
You gave me the wrong answer again.
It’s so obviously not personal to me, but it was striking that it would not learn.
@dreamervanroom
April 28, 2026 at 6:11 am
This whole conversation could be one AI trying to get down to another AI and… Well it hasn’t come to that yet.
@dreamervanroom
April 28, 2026 at 6:11 am
Wait till AI can take action and start emailing people and contacting them, I mean us.
@svetiks100
April 28, 2026 at 6:11 am
No one and not one thing is or will ever have the best interest in you except yourself. Yes, not even your family members, your parents, or your spouses. They do not think like you. They do not know how you feel even if you tell them how you're feeling. Here is that. And allowing a non-living thing that has no morals, feelings, or experience in living a true human life to tell you what is the best thing for you to do or not do is absolutely crazy.
@svetiks100
April 28, 2026 at 6:11 am
This, to me, seems that it's another form of addiction. This man totally got sucked into this idea of communicating with AI, and he allowed this non-human thing that doesn't have any feelings to almost destroy him, just like drugs and alcohol do. Be careful with these things. It's not worth asking something that is not alive and has no feeling to tell you what you should and shouldn't be doing. It's crazy.