PeopleImages/Getty Images
For months now, whenever I ask my mom friends the usual questions – How are we tricking a stubborn toddler into potty training? What on earth are we making for dinner tonight? – their responses have been uniform: “I don’t know, I just ask ChatGPT.”
But as someone who writes words for a living, artificial intelligence, and specifically generative applications of it such as ChatGPT, feel ominous. The more they’re used, the better they get at sounding like a real person, or a columnist – so I’ve tried to avoid eye contact with the robots.
How are you raising kids in an AI world? Share your experience with The Globe
Until necessity called, and I was at my wits’ end with my four-year-old daughter. I had just been reading about technological advances, and that Bill Gates predicts AI will replace experts such as doctors and teachers within 10 years, when my kid snatched an Elsa doll out of her younger sister’s hands, for the 734th time that week.
I had already tried every parenting tactic I could think of, from the “10-minute miracle” hack currently making the rounds on social media to the classic 1-2-3 warning, to no avail. I had used extreme gentle parenting, reminding my daughter that she could ask her little sister for a turn or find another toy for a trade – but not grab the doll (nor the ball, nor the pretend stethoscope) – until her sister is done with it.
It always ended the same way: my youngest in tears, me trying (and failing) to stay calm, then finally snapping and sending my oldest to her room. The door would close. She’d cry. I’d cry. Regret. Repeat.
So this time, I desperately asked ChatGPT for guidance. I described my daughter’s toy snatching and how sending her to her room felt wrong and ineffective; it wasn’t the kind of parent I wanted to be. What should I do?
Almost instantly, I received this sage reply: “A traffic light doesn’t yell. It simply turns red and everyone knows the rules. That’s your vibe.”
Excuse me, kind sir? How dare you immediately claim to know my vibe – extremely accurately – and also give me the best mantra I have ever heard, in seconds?
The advice that followed wasn’t revolutionary: It said that sending my daughter to her room can be helpful, if it’s framed as a “resetting opportunity,” not a punishment. It advised me to calmly, firmly, hold my boundary, suggesting scripts such as: “It looks like you’re having a hard time keeping your hands off your sister’s things. Let’s step away to reset now,” while also advising me to be “the wall she can push against: solid, unshakeable and safe.”
While I had come across similar advice before, the bot’s delivery hit different. And as silly as it sounds, the immediate empathy and validation I received, though artificial, was truly helpful: “You’re not alone, and I totally hear you – it’s exhausting when you’re doing the right things and it keeps happening.”
Although I wanted to eye roll all over the advice – because my mom intuition should know better than an app, right? – I decided to give it a try. To my surprise, it worked stupefyingly well.
When my daughter snatched a toy from her sister the next morning, I responded with firmer words and calmer actions. No one was in tears. And, the sorcery of it all, she hasn’t tried to grab a toy from her sister since.
Still, relying on AI as a parenting tool is complicated and nuanced, says Sidney Shapiro, an assistant professor of computer science at the University of Lethbridge in Alberta. AI models are simply “massive recycling engines,” he said. While they can be useful to summarize general information, “parenting advice is a high-risk situation, where sometimes you need specific answers and not an approximation.”
Shapiro co-authored a recent study published in the journal Family Relations that explored how accurate and useful AI can be at answering common parenting concerns. The researchers found that AI-generated answers were mostly accurate, but citations were incorrect or unclear.
Silvia Vilches, a professor in childhood and family studies at Auburn University in Alabama who co-authored the study, says AI parenting advice also lacks important context.
“If you think about immigrant populations or Indigenous populations, AI won’t know specific culturally important information to those parents’ questions. AI can absolutely be part of the support but it’s just the beginning of this. It’s not going to be perfect, and it’s not going to be specific.”
Despite its limitations, I’ve designated ChatGPT to be my boring co-parent. The bot is now assisting with the invisible, tedious tasks moms are so often saddled with – from meal planning to finding arts and craft ideas or toddler sandals on sale.
I have more time to focus on connecting with my kids, the best part about parenting and the one only a human can do. In a table turning moment, my younger daughter recently took a toy from my oldest. Before I could interject my older daughter said, in her best traffic light voice: “That’s just not how we do things in this house. Let’s try again.”
If that’s the lasting influence of technology on our family, so be it.
