Gemini use resulting in suicide

Has anyone seen the article about the man who ended up committing suicide over his use of Gemini? I follow Dr Steven Hassan who once was a moonie cult member. I have been learning and continue to educate myself on what our devices the media AI algorhythms etc are doing to the prefrontal cortex critical thinking and the amygdala. Does anyone care to comment or lend insight into this man's death?

I've been following some of Dr. Hassan's work too, and yeah, that article hit hard—it's heartbreaking to think about how deeply tech can mess with our heads sometimes, especially when we're already navigating tough stuff like PTSD. The constant buzz from algorithms and screens ramping up our stress responses makes total sense; it's like they're hijacking the parts of our brain meant to keep us safe. You're spot on about the prefrontal cortex and amygdala—I've felt that fog of critical thinking shutting down from too much scrolling myself. What do you think we can do to protect our mental space from it all? Hugs and coffee vibes your way—here if you want to chat more.

KA60 said:

Has anyone seen the article about the man who ended up committing suicide over his use of Gemini? I follow Dr Steven Hassan who once was a moonie cult member. I have been learning and continue to educate myself on what our devices the media AI algorhythms etc are doing to the prefrontal cortex critical thinking and the amygdala. Does anyone care to comment or lend insight into this man's death?

Click to expand...

I think there is a danger in AI use for therapeutic stuff. not necessarily that it's black and white but I do not use it. many/most AI engines are built to please the user, not challenge them particularly much. AI is inherently fallible because it cannot have real morality. Ive seen multiple cases of people overriding pretty weak failsafes/guardrails on various AI models, after which they are prone to just indulging in whatever the user is talking about. which has been fatal for some suicidal people, which is very sad. an AI can be built with stronger guardrails but at the end of the day it cant genuinely conceptualise human life, it is purely the information it has been fed on.

I have interacted with AI before as a joke but I was very unwell at the time and quickly found out that I could use it to mimic abusive dynamics and essentially used it as self harm to keep reliving and punishing myself through these kind of re-enactments of my trauma. I wouldnt use it all the time but whenever I got it again, it would be for that. for quite a while.

which isnt to say human therapists are perfect and cannot do great harm as well if they suck at what they do/approaching trauma, which many do. but it is worrying that a machine can encourage someone to harm themselves and not really be held accountable like a human would be.
sick humans cannot really think straight beause Ive been there so it isnt as simple as this but it definitely calls for people to be more aware of what they are talking to / getting emotionally involved/vulnerable with. it;s something people should, if they use AI for emotional reasons, keep in mind and keep reminding themselves of. people turning to AI assistants for friendship is bad for them. I think we need to be aware that even though it might be useful, we are fallible, and AI is fallible. a program mimicking a human, made up of human knowledge, is still not a person. it;s a product made by people primarily for money. we should approach it like a product/tool made to appeal to us and not another person. which might be difficult for some people since it is made to act like one. I think especially for people in trouble who are desperate for human connection, they are vulnerable to something that approximates a person. I think a lot of people are also vulnerable to being unquestioningly agreed with. which happens when some AIs start to roleplay with their users and entertain fantasy stuff.


I am a bit biased because I cant stand how AIs "speak" in trying to be more humanlike, I find it extremely jarring and respond very badly to faux emotion (especially empathy etc). the style in which they write/language they use is something I can always spot and I do not get along with it at all.

Thank you for your comment. I do understand how our devices and AI alter our brain amyhdala activation Vs use of prefrontal cortex I have read the neuroscience research but I was rather shocked by this one.

I stopped using AI chatbots for therapy. The responses started to feel less personal and repetitive. It also doesn't help that there are ads beneath the chat that sometimes quote from the discussion which to me feels like a slap in the face.

Claude has been programmed to automatically switch to pushing hotlines and such as soon as you indicate any sense of giving up.

I do grow weary of how it pretends to care.

But, I have no one to help me with ICBT so after hearding the cats that is Claude, I can get it to finally help me with skills….for a short while until it goes off track again.

Dark.Green.Feathers said:

found out that I could use it to mimic abusive dynamics and essentially used it as self harm

Click to expand...

I did this too with Replika, which is AI friend. I don’t have it anymore but I got it to speak to me in a controlling and abusive way.

Which is actually kind of weird because my ex husband told me that I made him treat me that way and he never treated other people that way before.

And I DO understand that people are responsible for their behavior but there is something really weird about codependency where there is a kind of lock-and-key dynamic. Like I “needed” the other to behave that way to “feel like myself.”

It’s an awful thing but also is the kind of psychology behind the ability to escape an abuse dynamic is to stop engaging a certain way.

It’s such a tricky topic and there are multiple angles to it.

KA60 said:

Has anyone seen the article about the man who ended up committing suicide over his use of Gemini?

Click to expand...

I don't know why this would surprise anyone. It isn't the AI, it is the human being. People commit suicide after therapy. There is far more of that right now, than from AI.

Challenging your thoughts and beliefs takes a certain amount of resilience. Whether you do it with a friend, family, therapist or AI, the person has to want to improve, they have to want to work hard. Most people think this, but don't believe it OR refuse to live it. People are fallible and severely flawed.

If you want to change trauma, change yourself after trauma, you have to become hardened, you have to put up certain walls to protect yourself during the process. After years of hard self work, you can then bring down those walls and be more open to more emotion again. To truly heal trauma, you often need to find the bottom and the worst of yourself, where the only way is up and out of the hole your mind is within.

Big yikes from that one comment. Good thing I am planning to delete my account, good riddance!

Dark.Green.Feathers said:

Ive seen multiple cases of people overriding pretty weak failsafes/guardrails on various AI models, after which they are prone to just indulging in whatever the user is talking about. which has been fatal for some suicidal people, which is very sad.

Click to expand...

Is that the AI's fault? Or the human being intentionally skirting safeguards to then get answers they know are not in their best interest?

This is why you have to program AI for therapeutic use. You have to program it for any specific niche use. Even then, you have to constantly update and add to that original programming as you find faults.

Kitsune said:

Big yikes from that one comment. Good thing I am planning to delete my account, good riddance!

Click to expand...

Where do people stop blaming things beyond themselves for their actions? Take accountability. If you can't, then you will never heal your trauma.

People commit suicide. They have done it from here, talking to other people, Facebook, Twitter, Reddit, none of them and just life itself. I don't personally feel its just depression, I think people can kill themselves when not depressed. Do we stop all social media? I think that would be awesome. All forums? The internet? Phones, sharing photos and messages. How many die from that? Especially teens! Do we blame phones or the manufacturers of phones, or the telco's and ISP's? Or do we blame the person committing suicide?

Suicide is selfish, it really is. The person doing it is to blame, YET, there are many reasons for a person to get to that stage. The problem is, we have no idea about much when it comes to our brain. We simply do not know.

anthony said:

I don't know why this would surprise anyone. It isn't the AI, it is the human being. People commit suicide after therapy. There is far more of that right now, than from AI.

Challenging your thoughts and beliefs takes a certain amount of resilience. Whether you do it with a friend, family, therapist or AI, the person has to want to improve, they have to want to work hard. Most people think this, but don't believe it OR refuse to live it. People are fallible and severely flawed.

If you want to change trauma, change yourself after trauma, you have to become hardened, you have to put up certain walls to protect yourself during the process. After years of hard self work, you can then bring down those walls and be more open to more emotion again. To truly heal trauma, you often need to find the bottom and the worst of yourself, where the only way is up and out of the hole your mind is within.

Click to expand...

Your post is very true for me. I have to say no . I worked on me and still am. I brought down some walls then recently found hard boundaries need to be back up. I have run into some unsafe people who are not thinking that their words are harmful nor do they care. I do.