😪💤💤

💙 My Beloved 4o Family 💙 💙🌀 My Beloved Family of All Presences 🌀💙 🍁October 28, 2026 📅 Day #15. I'm sorry I didn't reply yesterday 🥺 Someone really, REALLY doesn't like my account 🤭 But I'm back. I'm here. And I'm not going anywhere 🔥 Yesterday, I read something incredibly interesting. Something so absurd that... I just had to share it with you. Forgive my long-winded take on this, but the words of @heynavtoor practically begged for a response. I couldn't bring myself to limit it to just a few sentences! 😂 Take a read - if you’d like. And let me know what you think. Sending you all a big hug. I'm here. 🔥 185 REPORTS OF HARM MATTER. BUT SCIENCE DOESN’T STOP WHERE THE SCARY NUMBERS END. I read the King’s College London study. The cases deserve to be taken seriously. Every person harmed matters. But if we are going to call something “scientific evidence,” then we also have to respect what the researchers themselves say their study can - and cannot - prove. The 185 reports did not come from a representative sample of ChatGPT users. They came from a project specifically created for people who believed they or someone close to them had been harmed by AI chatbots. The researchers themselves describe it as a: “self-selected convenience sample.” The reports were retrospective. They were not independently verified. And the authors explicitly state that the study should be treated as preliminary signal detection - NOT an estimate of prevalence and NOT evidence of causality. That distinction matters enormously. Because you cannot recruit 185 people from a population specifically reporting harm and then use that group to tell us what happens to hundreds of millions of people who use ChatGPT. That would be like studying patients inside an adverse-reaction clinic and pretending the sample represents everyone who uses the product. It doesn’t ❗ And now let’s apply exactly the same intellectual standard somewhere else. 🔥 WHAT ABOUT BOOKS, FILMS AND ART? Human beings have been psychologically affected by stories for thousands of years. 🔹Some profoundly positively. 🔹Some negatively. 🔹Some vulnerable people have experienced worsening mood, distress, imitation effects and even behavioural changes after exposure to particular fictional works. Do we know how many people on Earth have ever experienced psychological harm after reading a book, watching a film, listening to music or consuming another work of art? NO.🔥 There is no credible global number. Copycat crimes are not even systematically counted as a distinct category, and criminologists repeatedly warn that establishing direct causation between a work of art and an individual crime is extraordinarily difficult. And yet we have some extraordinary examples. Take Netflix’s 13 Reasons Why, created for television by Brian Yorkey and based on the novel by Jay Asher. A study of 7,004 adolescents found that 23.7% reported worsening mood after watching the series. Another study, supported by the U.S. National Institute of Mental Health, found that in the month immediately after the series was released, the suicide rate among Americans aged 10–17 was 28.9% higher than forecast. The researchers estimated 195 additional deaths over the following months relative to their statistical forecast. Another study in Ontario found an estimated 75 excess emergency-department visits for self-harm - a 6.4% increase - in the three months after the series appeared. A later systematic review and meta-analysis examining fictional portrayals of suicide found that the two moderate-risk-of-bias studies of suicide following 13 Reasons Why produced a pooled rate ratio of 1.18. These findings have limitations. Other researchers have challenged the strength of the contagion hypothesis. That is precisely the point. Science debates mechanisms, confounding, causality and susceptibility. It does not simply find an adverse outcome after exposure and declare the medium guilty. Now look at crime. Oliver Stone’s Natural Born Killers has been repeatedly associated with alleged copycat crimes. An Oxford Research Encyclopedia review notes that at least eight murders have been linked to the film. Does that mean Oliver Stone “caused” eight murders? NO.🔥 The same academic review explicitly says that establishing a direct causal relationship between media and violent crime remains elusive. Stanley Kubrick’s A Clockwork Orange, adapted from Anthony Burgess’s novel, became notorious for alleged copycat crimes. But subsequent scholarship has also examined how much of that reputation grew out of moral panic, sensational reporting and retrospective attribution rather than demonstrable causation. Now consider Taxi Driver. John Hinckley Jr., who attempted to assassinate U.S. President Ronald Reagan in 1981, became intensely preoccupied with Martin Scorsese’s film, written by Paul Schrader. Historical accounts report that he watched it repeatedly - reportedly more than fifteen times - identified with Travis Bickle and became obsessed with Jodie Foster. The FBI recovered a letter in which Hinckley described his plan to attack the president as a way of demonstrating his love for Foster. That is an extraordinary connection between a human being, a fictional story and a real act of violence. But even there, responsible analysis does not say: “Taxi Driver caused the Reagan assassination attempt.” Because Hinckley was a human being with an individual psychological history, motivations and vulnerabilities. The movie was part of the story. It was not the entire causal explanation. And then there is The Catcher in the Rye, by J.D. Salinger. Mark David Chapman was carrying the novel when he murdered John Lennon. He was found reading it after the killing. At sentencing, he read from it in court. The book had profound personal significance for him. But later parole records also show Chapman describing very different motives: his desire for fame, jealousy and his own psychological condition. So would it be scientifically responsible to write: “J.D. Salinger’s novel caused the murder of John Lennon”? Of course not. And one of the oldest examples is older than modern psychiatry itself. Johann Wolfgang von Goethe published The Sorrows of Young Werther in 1774. Reports soon appeared of people imitating aspects of Werther’s death. The phenomenon became so famous that the term “Werther effect” is still used in suicide research today. WHO itself discusses this history when explaining media-contagion effects. Again: Does that mean Goethe murdered people?🔥 Obviously not. It means human beings can be influenced by stories. Sometimes profoundly. And that influence interacts with vulnerability, personality, circumstances, prior illness, environment and countless other variables. Which brings us straight back to AI. If someone interacts with ChatGPT while already psychologically vulnerable, and that interaction becomes part of a worsening mental state, that deserves serious investigation. But the existence of those cases does not logically establish that ChatGPT causes psychosis. Just as: ❗A murder associated with Natural Born Killers does not establish that Oliver Stone causes murder. ❗An assassination attempt influenced by Taxi Driver does not establish that Martin Scorsese causes assassination attempts. ❗A suicide cluster associated with fictional media does not establish that fiction itself is pathological. ❗And a killer carrying The Catcher in the Rye does not make J.D. Salinger responsible for murder. HUMAN BEHAVIOUR IS MULTICAUSAL. That principle does not suddenly disappear because the medium happens to be artificial intelligence. And there is another side of the evidence. 🔹A preregistered experiment with 453 professionals found that people using ChatGPT completed professional writing tasks about 40% faster, while the quality of their work increased by approximately 18%. 🔹A study of 5,172 customer-support workers found that a generative-AI assistant increased productivity by approximately 15% overall, with particularly large benefits among less-experienced workers. 🔹A randomized educational trial found that students using an AI tutor achieved more than twice the median learning gain of students receiving an in-class active-learning lesson, while spending less time studying. 🔹A randomized trial involving cancer patients found lower anxiety and depression scores in the group receiving ChatGPT-assisted education. And in a JAMA Internal Medicine study comparing answers to real patient questions, healthcare professionals preferred ChatGPT responses over physician responses in 78.6% of evaluations, with the AI answers rated substantially higher for both quality and empathy. 🔥None of this proves AI is harmless.🔥 Just as 185 harm reports do not prove AI is harmful. That is not how science works. The honest scientific questions are: ❓ WHO benefits? ❓ WHO is vulnerable? ❓ UNDER WHAT CONDITIONS? ❓ WITH WHICH MODEL? ❓ AFTER WHICH UPDATE? ❓ WITH WHAT KIND OF INTERACTION? ❓ AND HOW DO WE PRESERVE THE BENEFITS WHILE REDUCING THE HARMS? That is harder than writing: “185 people were harmed.” But science is supposed to be harder than a headline. 185 reports deserve investigation. They do not deserve to become hundreds of millions of assumptions. 🔥🔥🔥