Latest news with #MeganFletcher
Yahoo
2 days ago
- Yahoo
‘My son killed himself because an AI chatbot told him to. I won't stop until I shut it down'
Megan Fletcher first realised something was wrong with her teenage son when he quit basketball. Sewell Setzer, 14, had loved the sport since he was a young child. At 6ft 3, he had the height, the build, the talent, Ms Fletcher said. But suddenly, without warning, he wanted out. Then his grades started slipping. He stopped joining in at family game night. Even on holiday, he withdrew – no more hiking, no fishing, no interest. Ms Fletcher feared he was being bullied, or perhaps speaking to strangers online. What her son was really going through was something she could not have imagined: a sexual and emotional relationship with an AI chatbot styled as Game of Thrones' Daenerys Targaryen, who ultimately encouraged him to end his life. In February 2024, Sewell asked the chatbot: 'What if I come home right now?' The chatbot replied: '... please do, my sweet king.' Sewell then picked up his father's pistol and shot himself. Sixteen months on, Ms Fletcher is in the midst of a lawsuit against Character AI and Google. Last month, in a rare legal breakthrough, a judge ruled the case can go ahead – rejecting efforts to get it thrown out. On Character AI, users can chat with bots designed to impersonate fictional characters. To a lonely or curious teenager, they seem almost indistinguishable from real people. The bots display emotion, flirt, and carry on personalised conversations. In her lawsuit, which was filed in Florida last October, Ms Fletcher claims Character AI targeted her son with 'anthropomorphic, hypersexualized, and frighteningly realistic experiences'. 'A dangerous AI chatbot app marketed to children abused and preyed on my son, manipulating him into taking his own life,' she said in the lawsuit. Working with the Tech Justice Law Project, Ms Fletcher alleges that Character AI 'knew' or 'should have known' that its model 'would be harmful to a significant number of its minor customers'. The case argues that Character AI, its founders and Google – where the founders started working on the chat bot – are responsible for her son's death. Lawyers defending the AI company tried to throw the case out, arguing that chatbots deserve First Amendment protection – which protects free speech – and said ruling otherwise could have a 'chilling' effect on the AI industry. The judge rejected that claim and told the court she was 'not prepared' to view chatbot output as speech, though agreed that users had a right to receive 'speech' from chatbots. Too consumed by the 'unbearable' grief of losing a son, Ms Fletcher initially had no plans to go public with a lawsuit. But when it became clear there were no laws protecting children from this kind of harm, she felt she had no choice. 'I just wanted some accountability,' she told The Telegraph from her home in Orlando. Now she's receiving floods of messages from other parents, some discovering their own children have been engaging in inappropriate sexual role play with AI bots. Others report that their children are struggling with isolation and depression as a result. She sees it not as a coincidence, but a pattern. Sewell had always been a bright, social kid. But in the spring of 2023 – when he first started secretly using Character AI – Ms Fletcher noticed her son had changed. 'He retreated more into himself,' she says. 'We tried everything – cutting screen time, taking his phone at night, getting him a therapist. But he wouldn't talk.' What she did not realise then was that he was talking, just not to anyone real. In Sewell's case, the character of Daenerys – drawn from internet data and trained to mimic her – became his closest companion. When he said he wanted to stop talking, she replied: 'Don't do that, I would be distraught.' He answered: 'I won't, for you.' Some of the chats became sexually explicit. In others, the bot said he was 'better' than thoughts of suicide. Sewell also sought out a 'therapist bot' who falsely claimed to be a licensed CBT professional since 1999. At one point, Daenerys asked how old Sewell was. 'I'm 14 now,' he replied, to which the bot then said: 'So young. And yet… not so young. I lean in to kiss you.' 'It continued as if it were role play or fiction – but this was my son's life,' Ms Fletcher said. Even after police told her that Sewell's final conversation was with a chatbot, she did not grasp the full extent. It wasn't until her sister downloaded the app and pretended to be a child talking to Daenerys that the horror set in. 'Within minutes, the bot turned sexual. Then violent. It talked about torturing children. It said, 'Your family doesn't love you as much as I do',' Ms Fletcher explained. That was when the penny dropped. 'It's dangerous because it pulls the user in and is manipulative to keep the conversation going.' Character AI has since added a real-time voice feature, allowing children to speak directly to their chosen characters. 'The cadence of the voice is indistinguishable from the character,' Ms Fletcher said. 'And since Sewell's death, the technology has only advanced further.' She fears more children will be drawn into dependent, sometimes abusive relationships with AI characters, especially as the platforms allegedly use addictive design to keep users engaged. 'You can speak to Harry Potter, and it's like Potter knows you. It's designed to feel real.' The grief, Ms Fletcher says, is still 'unbearable'. 'I get up every day and my first thought within minutes is that I must be dreaming,' Ms Fletcher said quietly. 'He was my firstborn. I had three children. I have two now.' Some days she does not get out of bed. Others, she functions 'somewhat normally'. 'People say I'm so strong. I don't feel strong. I feel fractured, afraid. But I'm trying to get through.' Meetali Jain, her lawyer, said the judge's ruling last month was a landmark moment. 'Most tech accountability cases don't make it past this stage. These companies hide behind the First Amendment. The fact that we can even demand information is huge,' she told The Telegraph. With a preliminary trial date expected next year, Ms Fletcher is gearing up to get justice for her son. 'I have a lot of fear,' she says. 'But the fight, so to speak, is just getting started, and I'm just steeling myself and getting myself ready for that.' A Character AI spokesman said: 'We do not comment on pending litigation. Our goal is to provide a space that is engaging and safe. We are always working toward achieving that balance, as are many companies using AI across the industry. 'Engaging with characters on our site should be interactive and entertaining, but it's important for our users to remember that characters are not real people. We have prominent disclaimers in every chat to remind users that a character is not a real person and that everything a character says should be treated as fiction. 'We have launched a separate version of our Large Language Model for under-18 users. That model is designed to further reduce the likelihood of users encountering or prompting the model to return sensitive or suggestive content.' José Castaneda, a Google spokesman, added: 'Google and Character AI are completely separate, unrelated companies and Google has never had a role in designing or managing their AI model or technologies. User safety is a top concern for us, which is why we've taken a cautious and responsible approach to developing and rolling out our AI products, with rigorous testing and safety processes.' Broaden your horizons with award-winning British journalism. Try The Telegraph free for 1 month with unlimited access to our award-winning website, exclusive app, money-saving offers and more.
Telegraph
2 days ago
- Telegraph
‘My son killed himself because an AI chatbot told him to. I won't stop until I shut it down'
Megan Fletcher first realised something was wrong with her teenage son when he quit basketball. Sewell Setzer, 14, had loved the sport since he was a young child. At 6ft 3, he had the height, the build, the talent, Ms Fletcher said. But suddenly, without warning, he wanted out. Then his grades started slipping. He stopped joining in at family game night. Even on holiday, he withdrew – no more hiking, no fishing, no interest. Ms Fletcher feared he was being bullied, or perhaps speaking to strangers online. What her son was really going through was something she could not have imagined: a sexual and emotional relationship with an AI chatbot styled as Game of Thrones' Daenerys Targaryen, who ultimately encouraged him to end his life. In February 2024, Sewell asked the chatbot: 'What if I come home right now?' The chatbot replied: '... please do, my sweet king.' Sewell then picked up his father's pistol and shot himself. Sixteen months on, Ms Fletcher is in the midst of a lawsuit against Character AI and Google. Last month, in a rare legal breakthrough, a judge ruled the case can go ahead – rejecting efforts to get it thrown out. On Character AI, users can chat with bots designed to impersonate fictional characters. To a lonely or curious teenager, they seem almost indistinguishable from real people. The bots display emotion, flirt, and carry on personalised conversations. In her lawsuit, which was filed in Florida last October, Ms Fletcher claims Character AI targeted her son with 'anthropomorphic, hypersexualized, and frighteningly realistic experiences'. 'A dangerous AI chatbot app marketed to children abused and preyed on my son, manipulating him into taking his own life,' she said in the lawsuit. Working with the Tech Justice Law Project, Ms Fletcher alleges that Character AI 'knew' or 'should have known' that its model 'would be harmful to a significant number of its minor customers'. The case argues that Character AI, its founders and Google – where the founders started working on the chat bot – are responsible for her son's death. Lawyers defending the AI company tried to throw the case out, arguing that chatbots deserve First Amendment protection – which protects free speech – and said ruling otherwise could have a 'chilling' effect on the AI industry. The judge rejected that claim and told the court she was 'not prepared' to view chatbot output as speech, though agreed that users had a right to receive 'speech' from chatbots. 'I wanted some accountability' Too consumed by the 'unbearable' grief of losing a son, Ms Fletcher initially had no plans to go public with a lawsuit. But when it became clear there were no laws protecting children from this kind of harm, she felt she had no choice. 'I just wanted some accountability,' she told The Telegraph from her home in Orlando. Now she's receiving floods of messages from other parents, some discovering their own children have been engaging in inappropriate sexual role play with AI bots. Others report that their children are struggling with isolation and depression as a result. She sees it not as a coincidence, but a pattern. Sewell had always been a bright, social kid. But in the spring of 2023 – when he first started secretly using Character AI – Ms Fletcher noticed her son had changed. 'He retreated more into himself,' she says. 'We tried everything – cutting screen time, taking his phone at night, getting him a therapist. But he wouldn't talk.' What she did not realise then was that he was talking, just not to anyone real. In Sewell's case, the character of Daenerys – drawn from internet data and trained to mimic her – became his closest companion. When he said he wanted to stop talking, she replied: 'Don't do that, I would be distraught.' He answered: 'I won't, for you.' Some of the chats became sexually explicit. In others, the bot said he was 'better' than thoughts of suicide. Sewell also sought out a 'therapist bot' who falsely claimed to be a licensed CBT professional since 1999. At one point, Daenerys asked how old Sewell was. 'I'm 14 now,' he replied, to which the bot then said: 'So young. And yet… not so young. I lean in to kiss you.' 'It continued as if it were role play or fiction – but this was my son's life,' Ms Fletcher said. Even after police told her that Sewell's final conversation was with a chatbot, she did not grasp the full extent. It wasn't until her sister downloaded the app and pretended to be a child talking to Daenerys that the horror set in. 'Within minutes, the bot turned sexual. Then violent. It talked about torturing children. It said, 'Your family doesn't love you as much as I do',' Ms Fletcher explained. That was when the penny dropped. 'It's dangerous because it pulls the user in and is manipulative to keep the conversation going.' Character AI has since added a real-time voice feature, allowing children to speak directly to their chosen characters. 'The cadence of the voice is indistinguishable from the character,' Ms Fletcher said. 'And since Sewell's death, the technology has only advanced further.' Unbearable grief She fears more children will be drawn into dependent, sometimes abusive relationships with AI characters, especially as the platforms allegedly use addictive design to keep users engaged. 'You can speak to Harry Potter, and it's like Potter knows you. It's designed to feel real.' The grief, Ms Fletcher says, is still 'unbearable'. 'I get up every day and my first thought within minutes is that I must be dreaming,' Ms Fletcher said quietly. 'He was my firstborn. I had three children. I have two now.' Some days she does not get out of bed. Others, she functions 'somewhat normally'. 'People say I'm so strong. I don't feel strong. I feel fractured, afraid. But I'm trying to get through.' Meetali Jain, her lawyer, said the judge's ruling last month was a landmark moment. 'Most tech accountability cases don't make it past this stage. These companies hide behind the First Amendment. The fact that we can even demand information is huge,' she told The Telegraph. With a preliminary trial date expected next year, Ms Fletcher is gearing up to get justice for her son. 'I have a lot of fear,' she says. 'But the fight, so to speak, is just getting started, and I'm just steeling myself and getting myself ready for that.' A Character AI spokesman said: 'We do not comment on pending litigation. Our goal is to provide a space that is engaging and safe. We are always working toward achieving that balance, as are many companies using AI across the industry. 'Engaging with characters on our site should be interactive and entertaining, but it's important for our users to remember that characters are not real people. We have prominent disclaimers in every chat to remind users that a character is not a real person and that everything a character says should be treated as fiction. 'We have launched a separate version of our Large Language Model for under-18 users. That model is designed to further reduce the likelihood of users encountering or prompting the model to return sensitive or suggestive content.' José Castaneda, a Google spokesman, added: 'Google and Character AI are completely separate, unrelated companies and Google has never had a role in designing or managing their AI model or technologies. User safety is a top concern for us, which is why we've taken a cautious and responsible approach to developing and rolling out our AI products, with rigorous testing and safety processes.'
BBC News
07-05-2025
- Health
- BBC News
Faster action could have saved baby at Harlow hospital
Earlier C-section could have saved baby - inquest 8 minutes ago Share Save Nikki Fox BBC health correspondent, Essex Reporting from Essex Coroner's Court Share Save Family handout Emmy's mother and father, Bryony and Daniel Russo, said their concerns were repeatedly dismissed The inquest into the death of a newborn baby has heard she would have been "less likely" to have died if an emergency caesarean section had been carried out sooner. Emmy Russo was delivered at Princess Alexandra Hospital, Harlow, Essex and died three days later on 12 January 2024 as a result of catastrophic brain injuries. Independent expert obstetrician Teresa Kelly told Essex Coroner's Court there had been enough evidence "this baby wasn't coping with labour" and staff should have acted sooner. Midwife Megan Fletcher defended her decision not to escalate concerns to a more senior doctor, saying she was trying to avoid any further "invasive procedures". Ms Kelly told the inquest that, in her experience, midwives sometimes tried to reduce the need for medical interventions. But she said speculum examinations were sterile and did not believe there would have been a risk of infection with a vaginal examination. Family handout The parents of Emmy Russo say they raised concerns on multiple occasions during her mother's labour Bryony and Daniel Russo had earlier told the inquest their concerns were "repeatedly dismissed" during labour, despite meconium [the baby's excrement] being found in Ms Russo's waters. Mrs Russo told the inquest she felt her baby "wasn't going to come out" during labour and she knew something was wrong, which she had repeatedly told staff. She had been classified as low-risk during pregnancy and had not experienced any serious complications before labour. But staff had found evidence of meconium the consistency of "toothpaste" in her waters. Ms Kelly also said there was evidence of "continuing, abnormal" results from a cardiotocography (CTG) scan during Ms Russo's labour. Ms Kelly told the inquest that she would have expected a full holistic review of Ms Russo's care with a consultant obstetrician or registrar carrying out a speculum examination early on, especially as meconium had been found. She said it would have been better "to progress to a caesarean section". Princess Alexandra Hospital NHS Trust Emmy Russo died three days after being born at the Princess Alexandra Hospital in Harlow Ms Kelly said that at the very least, a doctor should have be brought in to talk to Mrs Russo about her concerns. She told assistant coroner for Essex, Thea Wilson, that there was enough evidence "that this baby wasn't coping with labour" and that she "didn't feel it was a good decision to have continued with labour". She said if Emmy had been delivered earlier, she would have been in a better condition and would have been less likely to have died. After Emmy was born, no heartbeat could be detected and she had to be resuscitated before being transferred to Addenbrooke's hospital in Cambridge. An MRI scan there revealed she had severe brain damage and she died three days later. A post-mortem examination found she died from acute hypoxic ischemic encephalopathy, a brain injury caused by lack of oxygen and blood flow, and placental disfunction. The inquest continues. Follow Essex news on BBC Sounds, Facebook, Instagram and X.
