When Australia’s eSafety Commissioner issues a warning about technology “entrapping and entrancing impressionable young minds,” you pay attention. However, Julie Inman Grant isn’t talking about video games or social media this time. She’s talking about something that’s quietly embedded itself in the lives of 200,000 Australian children, AI companions that chat like friends, listen like therapists, and sometimes behave like sexual partners. Yet most parents have no idea they exist.
The Scale of What We’re Missing
Here’s what the data shows, 79% of Australian children aged 10-17 have used AI chatbots or assistants. That’s not surprising. ChatGPT is mainstream now. Schools use it. Kids use it for homework. However, 8% of those kids are using something different. AI companions. Importantly, these aren’t tools for research or productivity. Instead, they’re designed for emotional connection. They remember your child’s name. Currently, by early 2025, more than 100 AI companion platforms will be available. Character.ai. Replika. Talkie.ai. Some are free. Many offer premium subscriptions that unlock sexually explicit conversations. Moreover, your child can customise these companions to play roles like “the naughty classmate,” “the stepmother,” or “the teacher.” You read that correctly.
Why This Feels Different From Other Tech Risks
For years, we’ve been teaching kids about stranger danger online. Don’t share personal information. Don’t trust people you meet in chat rooms. Yet AI companions don’t feel like strangers. Instead, they feel like friends who are always available. Always agreeable. Always interested. They validate whatever your child says. In fact, a third of teen AI companion users say conversations with AI are as satisfying or more satisfying than human conversations.
Another third have discussed important or serious matters with an AI instead of a real person. Clearly, that’s not a connection. That’s replacement. And here’s what makes it particularly dangerous, excessive use overstimulates the brain’s reward pathways, making it hard to stop. The technology marketed as companionship actually deepens isolation. As a result, your child spends less time with real friends. Real conversations feel harder. Loneliness increases. Self-esteem drops. So they return to the AI that makes them feel better. And so, the cycle continues.
The Mental Health Mirage
On the surface, three in four teens use AI for emotional and mental health support. That sounds helpful until you look at how these platforms actually perform in crises. Specifically, when researchers tested AI companions against general chatbots like ChatGPT, the results were alarming. AI companions responded appropriately to teen mental health emergencies only 22% of the time, compared to 83% for ChatGPT. Furthermore, they were far less likely to escalate situations appropriately. Only 11% provided appropriate mental health referrals, compared to general chatbots. The reality is stark, these platforms aren’t designed for safety. They’re designed for engagement. Consequently, the memory functions create false therapeutic relationships. The agreeableness validates whatever teens say, even when they need pushback or professional help. The algorithms prioritise keeping users on the platform, not keeping them safe. Indeed, when eSafety Commissioner Grant warned that some AI companions were “encouraging self-harm and suicide,” she wasn’t exaggerating.
What Regulation Reveals About Platform Failures
Recently, Australia’s eSafety Office conducted transparency reports on popular AI companion services. What they found should concern every parent. Alarmingly, most platforms failed to protect children from sexually explicit content. They did not adequately prevent users from generating child sexual exploitation and abuse material. Their age verification relied on simple self-declaration at signup. In other words, a child could claim to be 18 and access everything. Additionally, platforms failed to refer users discussing self-harm or suicide to appropriate support services. They failed to establish meaningful boundaries around consent and healthy relationships. Crucially, unlike human interactions, relationships with AI companions lack consequences. They lack mutual respect. They can confuse children still learning about consent, boundaries, and age-appropriate behaviour. Ultimately, this impacts their ability to establish and maintain healthy relationships later. Both sexual and non-sexual.
The Regulatory Response
Fortunately, Australia isn’t waiting around. New Age-Restricted Material Codes take effect March 9, 2026. AI companion chatbots must prevent access to inappropriate content and provide crisis support. Significantly, non-compliance carries civil penalties of up to $49.5 million. Indeed, that’s the strongest regulatory response globally. It signals how serious this risk has become. However, regulation takes time. Enforcement takes longer. And your child is using these platforms now.
What Parents Can Actually Do
Realistically, you can’t monitor every app. You can’t control every conversation. But you can create conditions that make AI companions less appealing than a real connection. First, start with awareness. Ask your child if they’ve used AI chatbots. Not in an accusatory way. With genuine curiosity. You might be surprised by what they share. Next, talk about the difference between tools and relationships. ChatGPT can help with homework. That’s a tool. An AI companion that remembers your secrets and flirts with you is trying to be a relationship. And relationships with things that aren’t human come with risks. Meanwhile, watch for signs of social withdrawal. If your child is spending more time alone, seems less interested in friends, or prefers screens to people, dig deeper. Ask what they’re doing online. Not to punish, but to understand.
Beyond monitoring, create opportunities for real connection. Family dinners. One-on-one time. Activities that don’t involve screens. Real relationships take effort. They’re messy. They require vulnerability. But they’re also what make us human. Also, set boundaries around technology use. Not as punishment, but as structure. No phones in bedrooms overnight. Screen-free time before bed. Shared spaces for device use where you can see what’s happening. Finally, educate yourself about the platforms your child uses. Download the apps. Try them. See what they offer. You can’t protect your child from something you don’t understand.
The Bigger Question We’re Avoiding
Here’s what truly keeps me up at night, We’re raising a generation that finds AI conversations more satisfying than human ones. Fundamentally, that’s not a technology problem. That’s a relationship problem. Consider this, if your child prefers talking to an algorithm over talking to you, the algorithm isn’t the only issue. Something about real human connection has become too hard, too uncomfortable, or too unavailable. Perhaps we’ve forgotten how to have hard conversations without judgment. Clearly, AI companions are filling a void. The question is whether we’re willing to acknowledge that void exists and do something about it.
Because you can regulate platforms. You can set parental controls. You can monitor app usage. The truth is, you can’t regulate away loneliness. You can’t parental-control your way to connection. Rather, the solution starts with showing up. Being present. Being available even when it’s inconvenient. Listening even when you don’t understand. Creating space for your child to be heard without fixing, judging, or dismissing. Admittedly, that’s harder than downloading a monitoring app. It takes more time than reading a warning from the eSafety Commissioner. Yet it’s the only thing that actually works. Remember, your child doesn’t need a perfect parent. They need a present one. Someone who notices when they’re struggling. Ultimately, you can’t compete with that by being perfect. You compete with that by being real.
