On Tuesday, OpenAI announced “Chat GPT for Teens,” a new feature designed, in the company’s words, to offer “stronger built-in safety protections for teens, including features to promote healthy use and additional controls for parents.” If ChatGPT’s systems suspect you’re a teenager, you have these restrictions until you prove otherwise. Very much how Roblox handled its (mostly successful?) chat restrictions.
AI’s problems are paramount, overwhelming, and often existential.
AI will be the generational technological diving line in years to come. Even if the “bubble” pops, it’s likely to be part of our children’s future. In what form? With what safety measures? With any respect for the material it steals from? Is there a reason we’re following the same “our platform is unsafe, time for teen mode” playbook?
The questions around “screen time” will soon extend, if they haven’t already, to AI.
The reason to want safer AI platforms is the same reason everyone should demand that Roblox is safer, too. You may not want—or let—your child to use either one, but the world’s better if we’re extracting higher standards. And what is the chatbot telling your child when you’re looking away, when they’re poking at a non-family device?
“Every generation of parents face some sort of technology they feel ill-equipped to judge,” said child psychologist Dr. Kathryn Steele, who’s also the director of clinical services at VocoVision, a company that provides teletherapy services for students.
I struggle with the AI question because trying to bury the concept of AI from my children forever doesn’t feel right, but at their current ages—9, 6—I fail to buy the arguments for why they should get access to it, either. I don’t know “when” the right age is to err on the side of pragmatism…but it’s one of those “I’ll know when I know.”
Or, more likely, when my oldest asks me to permission to use one of these platforms, and then we have a conversation about why. So far, it hasn’t come up. That’ll change.
(Some versions of tech shortcuts end up impossible to avoid, like my nine-year-old asking a devices in the house to do basic math problems instead of using her brain.)
“Parents assume that because they don’t understand how the tool works that they can’t judge if it’s safe,” said Steele. “But that’s not totally true. The key is noticing how your child is reacting to the tool or platform. If the tool they’re using is causing negative behavior like lack of sleep, acting out, or general emotional distancing, it’s clear that the tool is harming more than helping.”
That advice tracks with my own personal experience. When a parent asks me “How much screen time is appropriate?” I respond with “How does your child respond to screen time being taken away? “What are their emotions a device has to be put down?”
It’s one thing to read an academic paper, or read through what’s scientifically recommended as guidance, but your child’s behavior is the part that really matters.
One of my first meaningful interactions with AI was asking ChatGPT to add up my expenses and discovering it couldn’t do basic math. Instead of telling me that it couldn’t do math, though, it…basically…guessed? I wanted to scream and laugh at the same time! Rather than admit a weakness, ChatGPT just barreled ahead with lies, which is precisely the kind of stuff you’d be worried about with a kid using it, too.
(It ended up being off by a few cents. A trivial mistake for my bookkeeping, but one that takes on greater stakes when you’re off by a few cents on much bigger matters.)
“I have used AI tools for various reasons both personal and professional,” said Steele. “A lot of the worries I’m seeing are just how addictive using these tools can be. I’ve even caught myself preferring the quick, predictable outputs of an AI prompt over the more time consuming do-it-yourself process. If my adult brain can subtly drift towards using a tool that minimizes effort, a child, whose brain isn’t fully developed yet, can fall more dependent and vulnerable to it.”
“A chatbot can feel harmless because it acts like a mirror that only reflects back what you want to see. It can be affirming to the person looking for answers or advice when it’s telling them what they want to hear. But real growth requires real world friction.”
That’s my main concern, as well. Short-circuiting basic logic, reasoning, and problem solving for a shortcut. The moment MapQuest came around, allowing me to print out directions, my brain jettisoned all sense of direction and I’ve never gotten it back!
“A chatbot can feel harmless because it acts like a mirror that only reflects back what you want to see,” said Steele. “It can be affirming to the person looking for answers or advice when it’s telling them what they want to hear. But real growth requires real world friction. The harm comes in when people, especially kids, stop using AI as just a tool and becomes a crutch they need to navigate or hide from real world.”
Perhaps it’s the kind of hyperbole and anxiety that confronts the early days of any new technology, but it’s also true that the transformative powers of this technology has been forcibly imprinted onto society as a wealth-seeking exercise more than personal empowerment. We don’t know the risks, because the companies who’ve been developing the technologies haven’t cared to ask the questions, and they only implement features like “ChatGPT for Teens” once lawsuits are threatening profits.
For now—for me, anyway—it’s not a concern. My kids do not have conversations with the devices around the house because none of them are equipped for it. They can play music, they can ask what the weather is. Beyond that, they aren’t helpful. At some point, that’s going to change. Will I overhear my kid talking to a glowing puck, in-between asking if they could play the latest song from Taylor Swift’s new album?
“For parents,” said Steele, “knowing when your child is over relying on a chatbot for these feelings is essential to help them not fall into a false-mirror trap of advice.”
I just hope I’ll be able to spot it when it happens.
Have a story idea? Want to share a tip? Got a funny parenting story? Drop Patrick an email.
Also:
The big shift coming for me is the new “Siri AI,” which means all the devices in the house will suddenly become more ChatGPT-like. That might prompt a shift.
Fortunately, I guess, my kids do not use Siri at all? My nine-year-old uses her Apple Watch to send text messages, but I’m not sure she’s going to chat with it.
My biggest worry is what schools do with AI. I suspect middle school, which for me is one year away, is where the rubber is going to meet the road on all of us.
I was appalled when my child (10) told me that his classmates were able ask AI questions for a class project on a school device. The same school asked parents to sign a form in the first year of school (4 or 5 years old in the UK) to say that parents have explained all possible dangers of the internet to our children and our children won't do anything they shouldn't online. I actually like the school and on the whole I think they do their best - but their use of technology in teaching is extremely shoddy. I absolutely do not trust them to either teach my child about AI or to prevent him from accessing it. My kid isn't especially interested in AI but like all humans he likes to take the easy way out.