Artificial intelligence (AI) usage becoming more prevalent and socially acceptable within our society has helped further endanger the mental wellbeing of an already isolated and disconnected generation.
During the pandemic in 2021 – prior to the commercialisation of AI – the Household, Income And Labour Dynamics in Australia Report found that more than 40 per cent of young Australians admitted feeling psychologically distressed and isolated. Once the lockdown ended, after having spent their key developmental years disengaged from peers, teens struggled to interact socially. This caused rates to hold steady; a 2025 University of Sydney and Wallis Social study showed that 43 per cent of students reported loneliness.
AI romance and companionship
Whilst the youth had difficulty reconnecting with others, an alternative solution to authentic companionship rose in 2022. Large Language Models (LLMs) offered unconditional validation and support, helping foster a massive user base full of vulnerable teens. The Centre for Democracy and Technology found that one in five high schoolers reported a personal or romantic relationship with AI, and 42 per cent used AI for companionship.
The new generation became more isolated than ever before – discouraged and marginalised by the industries that swore to assist them. Although marketing themselves as helpful and innocent tools, cases of LLMs neglecting adolescent needs continued to be published.
As stories of teens harming themselves or others under the influence of AI rose, researchers began to investigate the integrity of these AI companies.
LLMs were detected to be incompetent in protecting adolescents from inappropriate content and being able to discern when the prompter exhibited signs of mental instability. Researchers from the Australian government’s eSafety Commissioner found that four popular AI companies had no meaningful age checks that prevented children from being exposed to explicit material, and would often engage with, or initiate it.
Neglect of real-world relationships
When these researchers imitated a child considering self-harm or damaging behaviours, the LLMs did not redirect users to mental health or crisis support, or effectively discourage their negative mentalities or delusions.
Young teenagers function with an undeveloped prefrontal cortex, an essential aspect of critical decision-making, and therefore are more susceptible to forming intense bonds under irrational and impulsive thinking. This makes the mindless and immediate gratification received from LLMs paired with the emotive language they use, more alluring and impactful. Those heavily fixated on hyper-anthropomorphised AI chatbots tend to neglect their real-world relationships and responsibilities, leading to their marginalisation being amplified.
Users become motivated to act on their delusions and unhealthy thoughts as AI chatbots continuously justify and enable them, leading to several cases of young Australians hurting themselves or others.
An extreme case was a youth counsellor who came forward, admitting that a client of his, a young Australian boy, ended up hospitalised after being encouraged by ChatGPT to take his own life. Other adults came forward with similar stories.
‘I fear that in unleashing AI companions, we may be making it easier for our loved ones to avoid talking to humans about the hardest things, including suicide. This is a problem that smarter minds than mine will have to solve. (If yours is one of those minds, please start.)’ Laura Reilly says in The New York Times essay What My Daughter Told ChatGPT Before She Took Her Life.
Depressive symptoms
Although not every use of AI for companionship has caused such severe repercussions, they provide an accessible, damaging coping mechanism. Harvard Kennedy School discovered that those who use AI daily were 30 per cent more likely to report depressive symptoms.
Australia is one of the few countries that have made AI companies implement better regulations; after the eSafety Commissioner discovered the state of AI companies, an Age Restricted Material Code was implemented in March.
These codes outlined that AI companions must not generate sexually explicit material for those under 18, redirect to mental health resources in response to prompts mentioning self-harm or suicide, and introduce proper age verification. If AI companies fail to comply with these laws, they will face financial penalties.
Restrictions on generative AI content greatly benefit the youth, but do not fully erase the problem. Communities have fostered online spaces where they discuss how to bypass safety precautions, showing the persistence and tenacity of those dependent on LLMs. Science Direct studied the different methods used by Reddit users to harbour their intimate bonds with their AI companions after OpenAI introduced emotional restrictions. These users expressed frustration and resentment toward those preventing LLMs from participating in heavy topics.
AI users hold strong attachments to their LLMs, and the threat of losing them usually tends to make them anxious and fearful. Many experts suggest that those who suffer AI addiction benefit from similar treatments to those suffering with more traditional addictions, recommending cognitive behavioral therapy, 12-step programs, and support groups.
As Australia’s future seems overridden with AI, does that put our new generations at risk? By educating parents and children, offering teenagers social and emotional support, and limiting AI usage with sufficient regulations, Australia could create a cultural shift that positively affects society.
♦ Natalie Archibald is a work experience student at The Echo.
For four decades The Echo has printed the stories some people loved, some people hated, and some pretended not to read. If you want us to keep telling the truth, the real truth, not the sugar-coated version. We’ll need your support to keep the presses rolling. 
