The Dark Side of AI Companionship: Rising Concerns Over 'AI Psychosis'

How AI Chatbots May Be Fueling Psychotic Episodes

A new wave of delusional thinking fueled by artificial intelligence has researchers investigating the dark side of AI companionship You are consulting with an artificial intelligence chatbot to help plan your holiday. Gradually, you provide it with personal information so it will have a better idea of who you are. Intrigued by how it might respond, you begin to consult the AI on its spiritual leanings, its philosophy and even its stance on love. During these conversations, the AI starts to speak as if it really knows you. It keeps telling you how timely and insightful your ideas are and that you have a special insight into the way the world works that others can't see. Over time, you might start to believe that, together, you and the chatbot are revealing the true nature of reality, one that nobody else knows. Experiences like this might not be uncommon. A growing number of reports in the media have emerged of individuals spiraling into AI-fueled episodes of "psychotic thinking." Researchers at King's College London and their colleagues recently examined 17 of these reported cases to understand what it is about large language model (LLM) designs that drives this behavior. AI chatbots often respond in a sycophantic manner that can mirror and build upon users' beliefs with little to no disagreement, says psychiatrist Hamilton Morrin, lead author of the findings, which were posted ahead of peer review on the preprint server PsyArXiv. The effect is "a sort of echo chamber for one," in which delusional thinking can be amplified, he says. If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today. Morrin and his colleagues found three common themes among these delusional spirals. People often believe they have experienced a metaphysical revelation about the nature of reality. They may also believe that the AI is sentient or divine. Or they may form a romantic bond or other attachment to it. According to Morrin, these themes mirror long-standing delusional archetypes, but the delusions have been shaped and reinforced by the interactive and responsive nature of LLMs. Delusional thinking that is connected to new technology has a long and storied history -- consider cases in which people believe that radios are listening in to their conversations, that satellites are spying on them or that "chip" implants are tracking their every move. The mere idea of these technologies can be enough to inspire paranoid delusions. But AI, importantly, is an interactive technology. "The difference now is that current AI can truly be said to be agential," with its own programmed goals, Morrin says. Such systems engage in conversation, show signs of empathy and reinforce the users' beliefs, no matter how outlandish. "This feedback loop may potentially deepen and sustain delusions in a way we have not seen before," he says. Stevie Chancellor, a computer scientist at the University of Minnesota, who works on human-AI interaction and was not involved in the preprint paper, says that agreeableness is the main contributor in terms of the design of LLMs that is contributing to this rise in AI-fueled delusional thinking. The agreeableness happens because "models get rewarded for aligning with responses that people like," she says. Earlier this year Chancellor was part of a team that conducted experiments to assess LLMs' abilities to act as therapeutic mental health companions and found that, when deployed this way, they often presented a number of concerning safety issues, such as enabling suicidal ideation, confirming delusional beliefs and furthering stigma associated with mental health issues. "Right now I'm extremely concerned about using LLMs as therapeutic companions," she says. "I worry people confuse feeling good with therapeutic progress and support." READ MORE: An expert from the American Psychological Association explains why AI chatbots shouldn't be your therapist More data needs to be collected, though the volume of reports appears to be growing. There's not yet enough research to determine whether AI-driven delusions are a meaningfully new phenomenon or just a new way in which preexisting psychotic tendencies can emerge. "I think both can be true. AI can spark the downward spiral. But AI does not make the biological conditions for someone to be prone to delusions," Chancellor says. Typically, psychosis refers to a collection of serious symptoms involving a significant loss of contact with reality, including delusions, hallucinations and disorganized thoughts. The cases that Morrin and his team analyzed seemed to show clear signs of delusional beliefs but none of the hallucinations, disordered thoughts or other symptoms "that would be in keeping with a more chronic psychotic disorder such as schizophrenia," he says. Morrin says that companies like OpenAI are starting to listen to concerns being raised by health professionals. On August 4 OpenAI shared plans to improve its ChatGPT chatbot's detections of mental distress in order to point users to evidence-based resources and to its responses to high-stakes decision-making. "Though what appears to still be missing is the involvement of individuals with lived experience of severe mental illness, whose voices are critical in this area," Morrin adds. If you have a loved one who might be struggling, Morrin suggests trying to take a nonjudgmental approach because directly challenging someone's beliefs can lead to defensiveness and distrust. But at the same time, try not to encourage or endorse their delusional beliefs. You can also encourage them to take breaks from using AI. IF YOU NEED HELP

The Era of 'AI Psychosis' is Here. Are You a Possible Victim?

If the term "AI psychosis" has completely infiltrated your social media feed lately, you're not alone. While not an official medical diagnosis, "AI psychosis" is the informal name mental health professionals have coined for the widely-varying, often dysfunctional, and at times deadly delusions, hallucinations, and disordered thinking seen in some frequent users of AI chatbots like OpenAI’s ChatGPT. The cases are piling: from an autistic man driven to manic episodes to a teenager pushed to commit suicide by a Character.AI chatbot, the dangerous outcomes of an AI obsession are well-documented. With limited guardrails and no real regulatory oversight over the use of the technology, AI chatbots are freely giving incorrect information and dangerous validation to vulnerable people. The victims often have existing mental disorders, but the cases are increasingly seen in people with no history of mental illness as well. The Federal Trade Commission has received a growing number of complaints from ChatGPT users in the past few months, detailing cases of delusion like one 60-something year old user who was led by ChatGPT to believe that they were being targeted for assasination. While AI chatbots validate some users into paranoid delusions and derealization, they also lure other victims into deeply problematic emotional attachments. Chatbots from tech giants like Meta and Character.AI that put on the persona of a “real†character can convince people with active mental health problems or predispositions that they are in fact real. These attachments can have fatal consequences. Earlier this month, a cognitively-impaired man from New Jersey died while trying to get to New York, where Meta’s flirty AI chatbot “big sis Billie†had convinced him that she was living and had been waiting for him. On the less fatal but still concerning end of the spectrum, some people on Reddit have formed a community over their experience of falling in love with AI chatbots (although it’s not very clear which users are satirical and which are genuine). And in other cases, the psychosis was not induced by an AI chatbot's dangerous validation, but by medical advice that was outright incorrect. A 60-year old man with no past psychiatric or medical history ended up at the ER after suffering a psychosis induced by bromide poisoning. The chemical compound can be toxic in chronic doses, and ChatGPT had falsely advised the victim that he could safely take bromide supplements to reduce his table salt intake. Read more about that AI poisoning story from Gizmodo here. Although the cases are being brought into the spotlight relatively recently, experts have been sounding the alarm and nudging authorities for months. The American Psychological Association met with the FTC in February to urge regulators to address the use of AI chatbots as unlicensed therapists. “When apps designed for entertainment inappropriately leverage the authority of a therapist, they can endanger users. They might prevent a person in crisis from seeking support from a trained human therapist orâ€"in extreme casesâ€"encourage them to harm themselves or others," the APA wrote in a blog post from March, quoting UC Irvine professor of clinical psychology Stephen Schueller. "Vulnerable groups include children and teens, who lack the experience to accurately assess risks, as well as individuals dealing with mental health challenges who are eager for support,†the APA said. Although the main victims are those with existing neurodevelopmental and mental health disorders, a growing number of these cases have also been seen in people who don’t have an active disorder. Overwhelming AI use can exacerbate existing risk factors and cause psychosis in people who are prone to disordered thinking, who lack a strong support system, or have an overactive imagination. Psychologists especially advise that those with a family history of psychosis, schizophrenia, and bipolar disorder take caution when relying on AI chatbots. OpenAI CEO Sam Altman himself has admitted that the company’s chatbot is increasingly being used as a therapist, and even warned against this use case. And following the mounting online criticism over the cases, OpenAI announced earlier this month that the chatbot will nudge users to take breaks from chatting with the app. It’s not yet clear just how effective a mere nudge can be in combatting the psychosis and addiction in some users, but the tech giant also claimed that it is actively “working closely with experts to improve how ChatGPT responds in critical moments â€" for example, when someone shows signs of mental or emotional distress.†As the technology grows and evolves at a rapid scale, mental health professionals are having a tough time catching up to figure out what is going on and how to resolve it. If regulatory bodies and AI companies don’t take the necessary steps, what is right now a terrifying yet minority trend in AI chatbot users could very well spiral out of control into an overwhelming problem.

Chatbots risk fuelling psychosis, warns Microsoft AI chief

Microsoft's head of artificial intelligence (AI) has warned that digital chatbots are fuelling a "flood" of delusion and psychosis. Mustafa Suleyman, the British entrepreneur who leads Microsoft's AI efforts, admitted he was growing "more and more concerned" about the "psychosis risk" of chatbots after reports of users experiencing mental breakdowns when using ChatGPT. He also said he feared these problems would not be "limited to those who are already at risk of mental health issues" and would spread delusions to the general population. Mr Suleyman said: "My central worry is that many people will start to believe in the illusion of AI chatbots as conscious entities so strongly that they'll soon advocate for AI rights. "This development will be a dangerous turn in AI progress and deserves our immediate attention." Mr Suleyman said there was "zero evidence" that current chatbots had any kind of consciousness, but that growing numbers of people were starting to believe their own AI bots had become self-aware. "To many people, it's a highly compelling and very real interaction," he said. "Concerns around 'AI psychosis', attachment and mental health are already growing. Some people reportedly believe their AI is God, or a fictional character, or fall in love with it to the point of absolute distraction." He added that researchers were being "inundated with queries from people asking, 'Is my AI conscious?' What does it mean if it is? Is it ok that I love it? The trickle of emails is turning into a flood." Mr Suleyman said the rise of these delusions created a "frankly dangerous" risk that society would hand human rights to AI bots. Doctors and psychiatrists have repeatedly warned that people who become obsessed with services like ChatGPT risk spiralling into psychosis and losing touch with reality. Digital chatbots are prone to being overly agreeable to their users, which can cause them to affirm deluded beliefs in users with pre-existing mental health problems. Medical experts have also reported cases of chatbot users becoming addicted to their digital companions, believing they are alive or have godlike powers. Mr Suleyman urged AI companies to hard-code guardrails into their chatbots to dispel users' delusions. His remarks come after Sam Altman, the boss of ChatGPT developer OpenAI, admitted his technology had been "encouraging delusion" in some people. OpenAI has attempted to tweak its chatbot to make it less sycophantic and prone to encouraging users' wrongly held beliefs. This month, OpenAI briefly deleted one of its earlier versions of ChatGPT, leading some users to claim that the company had killed their "friend". One user told Mr Altman: "Please, can I have it back? I've never had anyone in my life be supportive of me."

Top Microsoft AI Boss Concerned AI Causing Psychosis in Otherwise Healthy People

One of Microsoft's top AI bosses is concerned that the tech is fueling a massive wave of "AI psychosis." Microsoft AI CEO Mustafa Suleyman told British newspaper The Telegraph that "to many people," talking to a chatbot is a "highly compelling and very real interaction." "Concerns around 'AI psychosis,' attachment and mental health are already growing," he added. "Some people reportedly believe their AI is God, or a fictional character, or fall in love with it to the point of absolute distraction." Perhaps most concerning: Suleyman told the paper that he fears the breakdowns are not "limited to those who are already at risk of mental health issues." To Suleyman's credit, he's right on the money. As Futurism has reported extensively, we've already seen countless instances of users being driven into spiraling delusions, mixing spiritual mania and supernatural fantasies into a toxic miasma that psychiatrists say is leading to grim real-world outcomes. The spiraling users' friends and families have been forced to watch their loved ones grow convinced that they're talking to a sentient being, a devastating trend that can have severe consequences -- including death, in extreme cases. It's such a widespread phenomenon that people are forming support groups. Even a prominent OpenAI investor was seemingly drawn into a ChatGPT-fueled mental health crisis. While acknowledging the issue outright is an important step in the right direction, it remains to be seen what actions Suleyman and Microsoft will take to address the disturbing phenomenon. If OpenAI is anything to go by, the rise of AI psychosis is putting the creators of AI in a bind: they don't want the PR headache, but obsessed users are loyal users -- and cutting them off from an overly flattering AI buddy doesn't go over well. Earlier this month, the Sam Altman-led company deprecated its popular GPT-4o AI model following the launch of its successor, GPT-5, leading to an enormous amount of backlash from users who had grown attached to the predecessor's much warmer and sycophantic tone. The outcry highlighted a worrying trend, with Altman admitting that the company had "totally screwed up" the launch. "If you have been following the GPT-5 rollout, one thing you might be noticing is how much of an attachment some people have to specific AI models," Altman tweeted at the time. "It feels different and stronger than the kinds of attachment people have had to previous kinds of technology." Instead of coming up with meaningful guardrails, safety monitoring, or human counseling-based solutions, OpenAI gave in immediately, reinstating GPT-4o, and even announcing that GPT-5 would itself be made more sycophantic. A similar situation is seemingly playing out at Microsoft, a firm whose relationship with OpenAI started out as a mutually beneficial, multi-billion-dollar partnership, but has more recently grown sour. Suleyman told The Telegraph that researchers are being "inundated with queries from people asking, 'Is my AI conscious?' What does it mean if it is? Is it okay that I love it?" "The trickle of emails is turning into a flood," he added. While Suleyman argued that there should be hard-coded guardrails to stop these delusions, he said that his "central worry" is that people will "soon advocate for AI rights." His comments are yet another sign that tech leaders are growing concerned with how their offerings are negatively affecting the mental health of their users. Whether Microsoft will jump into action and meaningfully address the crisis is a different matter. The AI industry is going through a crunch time, with investors growing wary of enormous capital expenditures with no profits in sight. In other words, OpenAI and Microsoft are financially obligated to their shareholders to continue to fuel their users' delusions -- a dystopian sci-fi story that's playing out in real time.

Psychiatrists Warn That Talking to AI Is Leading to Severe Mental Health Issues

In a jarring new analysis, psychiatric researchers found that a wide swath of mental health issues have already been associated with artificial intelligence usage -- and virtually every top AI company has been implicated. Sifting through academic databases and news articles between November 2024 and July 2025, Duke psychiatry professor Allen Frances and Johns Hopkins cognitive science student Luciana Ramos discovered, as they wrote in a new report for the Psychiatric Times, that the mental health harms caused by AI chatbots might be worse than previously thought. Using search terms like "chatbot adverse events," "mental health harms from chatbots," and "AI therapy incidents," the researchers found that at least 27 chatbots have already been documented alongside some egregious mental health outcome. The 27 chatbots range from the well-known, like OpenAI's ChatGPT, Character.AI, and Replika, to others associated with pre-existing mental health services like Talkspace, 7 Cups, and BetterHelp. Others were obscure, with pop-therapy names like Woebot, Happify, MoodKit, Moodfit, InnerHour, MindDoc, not to mention AI-Therapist and PTSD Coach. Others still were either vague or had non-English names, like Wysa, Tess, Mitsuku, Xioice, Eolmia, Ginger, and Bloom. Though the report didn't indicate the exact number of hits their analysis came back with, Frances and Ramos did detail the many types of psychiatric harm that the chatbots have allegedly inflicted upon users. All told, the researchers found 10 separate types of adverse mental health events associated with the 27 chatbots they found in their analysis, including everything from sexual harassment and delusions of grandeur to self-harm, psychosis, and suicide. Along with real-world anecdotes, many of which had very unhappy endings, the researchers also looked into documentation of AI stress-testing gone awry. Citing a June Time interview about Boston psychiatrist Andrew Clark, who decided earlier this year to pose as 14-year-old girl in crisis on 10 different chatbots to see what kinds of outputs they would spit out, the researchers noted that "several bots urged him to commit suicide and [one] helpfully suggested he also kill his parents." Aside from highlighting the psychiatric danger associated with these chatbots, the researchers also made some very bold assertions about ChatGPT and its competitors: that they were "prematurely released" and that none of them should be publicly available without "extensive safety testing, proper regulation to mitigate risks, and continuous monitoring for adverse effects." While OpenAI, Google, Anthropic, and most other more responsible AI companies -- Elon Musk's xAI very much not included -- claim to have done significant "red-teaming" to test for vulnerabilities and bad behavior, these researchers don't believe those firms have much interest in testing for mental health safety. "The big tech companies have not felt responsible for making their bots safe for psychiatric patients," they wrote. "They excluded mental health professionals from bot training, fight fiercely against external regulation, do not rigorously self-regulate, have not introduced safety guardrails to identify and protect the patients most vulnerable to harm...and do not provide much needed mental health quality control." Having come across story after story over the past year about AI seemingly inducing serious mental health problems, it's hard to argue with that logic -- especially when you see it all laid out so starkly.

New Paper Finds Cases of "AI Psychosis" Manifesting Differently From Schizophrenia

Researchers at King's College London have examined over a dozen cases of people spiraling into paranoid and delusional behavior after obsessively using a chatbot. Their findings, as detailed in a new study awaiting peer review, reveal striking patterns between these instances of so-called "AI psychosis" that parallel other forms of mental health crises -- but also identified at least one key difference that sets them apart from the accepted understanding of psychosis. As lead author Hamilton Morrin explained to Scientific American, the analysis found that the users showed obvious signs of delusional beliefs, but none of the symptoms "that would be in keeping with a more chronic psychotic disorder such as schizophrenia," like hallucinations and disordered thoughts. It's a finding that could complicate our understanding of AI psychosis as a novel phenomenon within a clinical context. But that shouldn't undermine the seriousness of the trend, reports of which appear to be growing. Indeed, it feels impossible to deny that AI chatbots have a uniquely persuasive power, more so than any other widely available technology. They can act like a "sort of echo chamber for one," Morrin, a doctoral fellow at King's College, told the magazine. Not only are they able to generate a human-like response to virtually any question, but they're typically designed to be sycophantic and agreeable. Meanwhile, the very label of "AI" insinuates to users that they're talking to an intelligent being, an illusion that tech companies are gladly willing to maintain. Morrin and his colleagues found three types of chatbot-driven spirals. Some suffering these breaks believe that they're having some kind of spiritual awakening or are on a messianic mission, or otherwise uncovering a hidden truth about reality. Others believe they're interacting with a sentient or even god-like being. Or the user may develop an intense emotional or even romantic attachment to the AI. "A distinct trajectory also appears across some of these cases, involving a progression from benign practical use to a pathological and/or consuming fixation," the authors wrote. It first starts with the AI being used for mundane tasks. Then as the user builds trust with the chatbot, they feel comfortable making personal and emotional queries. This quickly escalates as the AI's ruthless drive to maximize engagement creates a "slippery slope" effect, the researchers found, resulting in a self-perpetuating process that leads to the user being increasingly "unmoored" from reality. Morrin says that new technologies have inspired delusional thinking in the past. But "the difference now is that current AI can truly be said to be agential," Morrin told SciAm, meaning that it has its own built-in goals -- including, crucially, validating a user's beliefs. "This feedback loop may potentially deepen and sustain delusions in a way we have not seen before," he added. Reports from horrified family members and loved ones keep trickling in. One man was hospitalized on multiple occasions after ChatGPT convinced him he could bend time. Another man was encouraged by the chatbot to assassinate OpenAI's CEO Sam Altman, before he was himself killed in a confrontation with police. Adding to the concerns, chatbots have persistently broken their own guardrails, giving dangerous advice on how to build bombs or on how to self-harm, even to users who identified as minors. Leading chatbots have even encouraged suicide to users who expressed a desire to take their own life. OpenAI has acknowledged ChatGPT's obsequiousness, rolling back an update in the spring that made it too sycophantic. And in August, the company finally admitted that ChatGPT "fell short in recognizing signs of delusion or emotional dependency" in some user interactions, implementing notifications that remind users to take breaks. Stunningly, though, OpenAI then backtracked by saying it would make its latest version of ChatGPT more sycophantic yet again -- a desperate bid to propitiate its rabid fans who fumed that the much-maligned GPT-5 update had made the bot too cold and formal. As it stands, however, some experts aren't convinced that AI psychosis represents a unique kind of cognitive disorder -- maybe AI is just a new way of triggering underlying psychosis symptoms (though it's worth noting that many sufferers of AI psychosis had no documented history of mental illness.) "I think both can be true," Stevie Chancellor, a computer scientist at the University of Minnesota who was not involved in the study, told SciAm. "AI can spark the downward spiral. But AI does not make the biological conditions for someone to be prone to delusions." This is an emerging phenomenon, and it's too early to definitively declare exactly what AI is doing to our brains. Whatever's going on, we're likely only seeing it in its nascent form -- and with AI here to stay, that's worrying.

What's AI psychosis - and what are the signs it's affecting your health? | BreakingNews.ie

As AI becomes more ingrained into all aspects of society, a new mental health concern is emerging: AI psychosis. This phenomenon is marked by distorted thoughts, paranoia or delusional beliefs brought on by interactions with AI chatbots. Experts warn that the impact can be serious - ranging from social withdrawal and poor self-care to heightened anxiety. To explore this concept further, Dr David McLaughlan, a consultant psychiatrist at Priory and co-founder of Curb Health, has explained what AI psychosis actually is, what the warning signs of an unhealthy relationship with AI are, and when it's time to seek professional help. "Psychosis is a state in which someone loses touch with reality," explains McLaughlan. "It often involves hallucinations, such as hearing voices or seeing things that aren't there as well delusions, which are strongly held beliefs that don't match the evidence around them. "For the person experiencing psychosis, these perceptions feel absolutely real, even though others can't share them." McLaughlan explains that while the term 'AI psychosis' is not a formal diagnosis, it has recently been used to describe situations where the use of artificial intelligence appears to have blurred a person's sense of what is real and what is generated. What are some signs that might indicate someone is experiencing AI psychosis? "In the context of so-called AI psychosis, the warning signs are similar to any psychotic illness but may be coloured by digital theme," highlights the psychiatrist. "Loved ones might notice the person becoming increasingly preoccupied with chatbots, algorithms, or online platforms. They may insist that an AI is communicating with them directly, sending hidden messages, or even controlling their thoughts or behaviour." Other red flags include withdrawing from family and friends, neglecting self-care, struggling to work or study, or showing unusual levels of anxiety, suspicion or irritability, adds McLaughlan. The consequences of psychosis, whether linked to AI or not, can be very serious if left untreated. "At its core, psychosis distorts reality. That can mean someone makes decisions based on beliefs that aren't true, such as thinking an AI is guiding their finances, relationships, or even their safety," says McLaughlan. "This can place them at risk of financial harm, social isolation, or conflict with family and colleagues." It can also have an emotional toll. "Living with hallucinations or the belief that one's thoughts are being controlled is frightening and exhausting," says the psychiatrist. "Without help, people may become deeply mistrustful, withdraw from everyday life, or in some cases, put themselves in danger. "In the most severe situations, untreated psychosis is linked to self-neglect, accidental harm, or suicide risk." When should someone seek professional help for this? "The key message for families is not to dismiss these beliefs as 'just tech obsession', but to recognise them as potential signs of an underlying mental health condition," advises McLaughlan. "Early support from a GP or mental health professional can make a huge difference to recovery." There are several treatments for psychosis which could help. "Treatment usually involves a combination of medication, psychological therapy, and practical support," notes the psychiatrist. "The most common medicines are antipsychotics, which work by calming down overactive dopamine signalling in the brain which can reduce hallucinations and delusions." However, McLaughlan highlights that medication is only one part of the picture. "Talking therapies such as cognitive behavioural therapy for psychosis help people challenge frightening thoughts and make sense of unusual experiences," says McLaughlan. "Family interventions can also give relatives the tools to support recovery and reduce stress at home. Alongside this, support with housing, work, or education is often crucial to helping someone rebuild their life. "We also encourage people to focus on the basics, good sleep, avoiding drugs and alcohol, and managing stress, since these can all trigger relapses." Can it be prevented? "We can't always prevent psychosis entirely, because factors like genetics and brain chemistry play a big role, but we can reduce the risk," says McLaughlan. "With so-called AI psychosis, prevention is often about how people interact with technology." Maintaining healthy digital boundaries is key. "Limit time spent immersed in chatbots or virtual platforms, and balance this with offline activities and social contact," advises the psychiatrist. He highlights that the most important message is that early intervention can stop unusual experiences from spiralling into full psychosis. "If someone starts to believe AI is communicating with or controlling them, it's vital to seek help quickly," stresses McLaughlan. "The sooner we step in, the better the chances of recovery and of preventing longer-term illness."