Dad Turns Roblox Into an AI‑Run ‘Truman Show’ to Shield His Son

It started as a late‑night rant on X, the kind of post that feels half‑joking, half‑serious, and then, boom, the internet latches on. Adam Lyttle, a self‑described "tech‑savvy dad," announced he had vibecoded a safer version of Roblox for his nine‑year‑old. The word “vibecoded” is his own invention, but the idea behind it is all too clear: replace the strangers his son might meet online with bots that never go off‑script.

"The game is populated with recordings of other playthroughs, giving the illusion of playing with others," Lyttle wrote, attaching a short clip of his custom Roblox map. In the video, virtual avatars sprint around a colorful obstacle course, chat bubbles pop up, and one of the faux‑players types, "im playing on my school chromebook rn," just like a typical teen on the platform. The twist? Those chat lines are generated on the fly by ChatGPT, meaning the AI pretends to be a real kid, a real friend, a real rival.

He went further, explaining that the bots even talk to each other. "He can talk with the bots and they chat back using ChatGPT," Lyttle added, clearly proud of the technical wizardry. He even posted a link to the private game, urging others to try it out. The tweet quickly snowballed into a frenzy of retweets, memes, and—unsurprisingly—outrage.

Critics were swift. "Willingly putting your own child into an AI‑generated simulation and lying to them about it is just wrong," one commenter wrote. Others likened the setup to the 1998 film The Truman Show, where the protagonist lives his entire life on a staged set, unaware that everyone around him is an actor. "This is dystopian and incredibly depressing," another user quipped, echoing the sentiment that a child’s trust could be eroded by such deception.

Beyond the moral arguments, many pointed out the sheer overkill of Lyttle’s solution. "You could just enable Roblox's parental controls, or pick a different game altogether," suggested a user who seemed exhausted by the drama. The consensus was clear: there are easier, less invasive ways to keep kids safe online.

When the backlash grew louder, Lyttle shifted his tone, framing the whole thing as a prank meant to generate engagement. "The tweet was perfectly constructed to be just rage‑y enough to drive engagement," he claimed, adding that he honestly didn't expect it to take off. Whether it was a stunt or a genuine attempt at protection, the episode adds another chapter to the growing saga of parents leaning on AI to raise their kids.

He's not alone in the AI‑parenting experiment. OpenAI CEO Sam Altman has openly talked about using ChatGPT to help with bedtime stories and even to generate personalized podcasts about his son’s day. Another father confessed to recording his toddler's babble and feeding it into a custom model that spits out a whole website dissecting every coo and giggle. The line between helpful tech and intrusive surveillance is getting fuzzier by the day.

What makes Lyttle’s case stand out is the scale of the illusion. He didn’t just give his son a chatbot to talk to; he built an entire fake multiplayer lobby, complete with AI‑driven avatars dodging obstacles, cheering, and dropping the occasional typo‑laden message. It’s a digital panopticon for a kid who thinks he’s just having fun with peers.

Some observers tried to inject a dose of practicality. "Coulda just let the kid play Mario instead of putting him in the ‘Truman Show,’" one witty comment read. Others wondered about the long‑term effects: Would the child later discover the truth and feel betrayed? Would the AI‑filled environment stunt his ability to navigate real‑world social cues?

At the end of the day, Lyttle’s experiment serves as a cautionary tale. Protecting children from online harm is vital, but the methods we choose say a lot about our values. Is a fabricated world with bots a safer haven, or does it become a gilded cage? The conversation is far from over, and as AI tools become more accessible, we’ll likely see more parents walking that fine line between innovation and overreach.