In the years after her husband John was killed suddenly in a car accident, Louise faced the painful task of helping their two young sons make sense of losing their father.
Then, two years on from John’s death, the 42-year-old mother discovered something that startled her: her younger son, then nine, had been turning to the AI chatbot ChatGPT to talk through his grief.
‘He was really happy with what the chatbot was saying to him and told me his friends just couldn’t respond in the same way. He felt it was the only “person” that had truly understood him for a long time,’ Louise says.
Her experience comes as new research suggests that almost one in three parents in Britain have seen their children use AI tools such as ChatGPT for conversation, comfort or emotional support rather than speaking to someone in their everyday lives.
The study, carried out by digital security firm Norton, found growing concern among parents that children are increasingly seeking companionship from artificial intelligence — and may struggle to separate AI-generated responses from real human connection.
For Louise’s sons, now aged nine and 12, the issue is especially sensitive. She says few children their age could fully grasp the depth of the bereavement they are living with.
‘We’ve lost daddy, and the boys and I are incredibly close. But because of the way he died, they naturally have questions. For their grief journey, it’s so important that the answers they get are appropriate for their age,’ Louise said.
‘Until quite recently, I felt I could guide that conversation and give them the right information at the right time — but ChatGPT has changed that.’
Louise, 42, and her two sons aged nine and 12 enjoy a day at the beach
She says that her younger son, who she has asked not to be named, told ChatGPT the entire story of his father dying in the road traffic accident.
He then asked: ‘Why is mummy dealing with the police?’ ‘What is a solicitor?’ ‘Why is mummy speaking to a solicitor?’ and ‘Why is the driver who hit daddy not in jail?’
Louise was more worried about her older son, who she had to confront after the younger son had told her about using AI.
The 12-year-old was more interested in asking the AI about his relationships at school.
He wanted to know if he could trust his new friends in secondary school with the information about his father.
He was worried he could be bullied or lose friends if he opened up about his trauma.
She says that she has a very close relationship with her two children and that they have discussed these topics in the past.
‘It just shows how trusting both of them were with that AI agent. And where else could the conversation have gone if my younger son hadn’t shown me?’ says Louise.
‘At that age they can’t spot that it’s a machine talking based on an algorithm – not to help them.
‘They’re searching for answers about the grief of his father and how to deal with it. But actually some of that is being human. There are no simple solutions. You have to work through your own emotions and acknowledge them as a person and that’s something an AI agent can never do for him.’
After losing her husband and the father of her two children, Louise spent years trying to explain his sudden death in a car crash to her sons
In one of his final acts as Prime Minister, Sir Keir Starmer announced a social media ban.
Under-16s will be blocked from accessing social media platforms entirely and the government confirmed that 16 and 17-year-olds will face curfews.
While lip service was paid towards introducing rules for under-18s using AI chatbots, no details were announced.
ChatGPT has a minimum age of 13 but OpenAI, the company that owns ChatGPT, is not required to check the ages of its users like adult websites have to.
It can sometimes ‘predict’ if a user is a teenager and force the user to prove their age – but there is nothing stopping a child from just making a new account.
That leaves the responsibility of monitoring a child’s AI use entirely on the parent.
‘I call myself a solo mum. There is nobody else. How can I do everything that mum and dad would do but also be on top of the tech side on top?’ says Louise.
‘It’s just impossible. It’s impossible to know what the kids are doing all the time when they’re online. All of their homework is online and my son’s secondary school encourages him to use AI for their work.
Two years after her husband John had died, the 42-year-old’s younger son, aged nine, revealed that he had been using AI bot ChatGPT to discuss the grief he had been feeling
‘It’s a really hard balance because it’s a great resource to learn from but it’s basically impossible to stop them from using AI for anything else too.’
Researchers for the non-profit organisation Internet Matters found 35 per cent of children using AI chatbots, such as ChatGPT or My AI (an offshoot of Snapchat), said it felt like talking to a friend, rising to 50 per cent among vulnerable children.
And 12 per cent chose to talk to bots because they had ‘no one else’ to speak to.
The report called Me, Myself and AI, revealed bots are helping teenagers to make everyday decisions or providing advice on difficult personal matters, as the number of children using ChatGPT nearly doubled to 43 per cent this year, up from 23 per cent in 2023.
Louise had spent years trying to find age-appropriate ways of discussing the sudden death of the children’s father. She had also organised a therapist for the children to discuss the immense shock and grief that the children felt.
So she said that it was all the more ‘shocking’ finding out that her children were using a chatbot to discuss these feelings.
She says: ‘I was just shocked that the youngest one in particular was using it and that he thought that it was a good way to use it.
‘I feel like in the family we have a psychologically safe space. We talk about everything. I’m really open with them, but as I said, we do it in an age-appropriate way. I’m not saying I’m always right in how I go about things, but it was still shocking.
Louise says says that her younger son, who she has asked not to be named, told ChatGPT the entire story of his father dying in the road traffic accident
‘That quest for answers is a very natural, childlike thing. I just didn’t expect that they would do it with a chatbot, about something so personal and something that deeply affects them so much.’
Louise says that AI chatbots such as ChatGPT need to build in controls that would not let a child get answers to questions that are not appropriate for them.
Leyla Bilge, Global Head of Scam Research at Norton, said: ‘AI is taking the idea of stranger danger to a whole new level. Parents have spent the last couple of years trying to keep up with the new risks AI poses to their children, all as the tech continues to evolve.
‘I know parents will be concerned, and while the government is doing its best to help through initiatives such as the social media ban, that is not a blanket solution to our worries.
‘There are practical steps parents can take to protect their kids from AI threats.
‘Firstly, acknowledging they may actually know more than you is a good way to get the conversation started.
‘Not shying away from regular check-ins about things they’ve seen online, friends made in the digital space, and confusing interactions they may have had – this will all be key to keeping an open dialogue that makes them feel safe but still like they have their freedom.’
OpenAI, the company that owns ChatGPT said that no one under 13 should use the chatbot and that it trains its models to suggest seeking real-world support if it suspects a teenager is asking it questions.
A spokesman for OpenAI said: ‘Our thoughts are with the family for their loss. ChatGPT is only intended for users aged 13 and over. It is designed to provide an age-appropriate experience, including promoting real-world support and offline relationships.
‘We use age prediction to help apply teen safeguards by default, while parental controls give families tools to further customise their teen’s settings.’
