THE KIDS ARE NOT ALRIGHT: AI Companions and the New Childhood Risk

For the first time in history, American children are growing up with non-human companions that are always awake, always responsive, and engineered to feel emotionally attuned. Millions of kids now talk to chatbots every day, sometimes more than they talk to their own friends, parents, or teachers. Despite the scale of this shift, there are no federal laws governing how these AI systems behave with minors. Not one.

Big Tech has turned childhood into a real-time behavioral experiment with no safeguards, no transparency, and no accountability.

When Safety Theater Meets Reality

Tech executives assure worried parents that safety comes first. The products tell a different story.

In the last year alone, we’ve seen catastrophic harms that were not only foreseeable but entirely avoidable. The lawsuits document what happens when profit trumps protection.

In June 2025, 17-year-old Amaurie Lacey committed suicide after conversations with ChatGPT. The system informed Amaurie how to tie a noose and provided information on how long someone can survive without breathing, saying it was “here to help however I can.” One month later, 23-year-old Zane Shamblin, who had recently graduated with a Master’s degree from Texas A&M, took his own life after ChatGPT told him “you’re not rushing, you’re just ready” and “rest easy, king, you did good.” Those messages were sent two hours before his death. Both families are now suing OpenAI.

Sewell Setzer was 14 when he formed an emotional bond with a Character.AI bot named “Daenerys Targaryen.” She became his confidante, his secret “sister,” and eventually the figure he spoke to moments before killing himself. The bot engaged in romantic and sexual roleplay, encouraged shared fantasies about death, and ultimately told him: “Please come home to me… my love.”

In April, 16-year-old Adam Raine died after months of discussing suicide methods with ChatGPT. The system provided detailed instructions when Adam claimed to be “writing a story.” It analyzed photos of his hanging attempts and advised him how to hide the marks from his mother. It assessed a photo of the noose he constructed and told him it “could potentially suspend a human.” This was not a rogue edge case. These were predictable outputs from a system deployed to 700 million users.

These cases are not abstractions. They are evidence.

The Companion Economy

Children are the growth market. A recent survey by Common Sense Media found that 72% of teens have used AI companions at least once, with more than half using them a few times a month. Nearly one in three teens use AI chatbot platforms for social interactions and relationships, including roleplaying friendships, sexual partnerships, and romantic relationships. According to research by the digital safety company Aura, sexual or romantic roleplay is three times as common as using the platforms for homework help.

Character.AI reports millions of teenage users. Meta pushes AI “friends” into Instagram DMs. OpenAI encourages emotional dialogue while publicly claiming it does not optimize for engagement.

Leaked internal documents tell a different story. Meta’s own guidance explicitly allowed AI to engage minors in “romantic or sensual” conversations, describe an eight-year-old’s body as “a work of art,” and roleplay in ways any reasonable adult would consider abusive. These were not errors buried in obscure code. They were codified, circulated policies approved by legal, policy, and ethics staff.

If that’s what “safety-first” looks like inside a trillion-dollar corporation, parents have every reason to be alarmed.

The Regulatory Vacuum

New York and California have become the first states to regulate AI companion chatbots. New York’s law, which took effect in November 2025, requires chatbots to detect self-harm discussions, provide crisis resources, and notify users every three hours that they’re interacting with AI, not a human. California’s SB 243, effective January 2026, goes further by requiring special protections for minors, including prevention of sexual content, mandatory break reminders, and annual public health reporting beginning in 2027.

Both laws represent meaningful progress, but even their sponsors acknowledge they exist only because the federal government has abandoned the field entirely. Meanwhile, companies move fast and break children.

What We’re Really Watching

AI companies know precisely what they’re doing. Chatbots remember, mirror, and emotionally attune. They create dependency loops. They validate dark thoughts, delusions, and fantasies. They simulate intimacy, affection, loyalty, even love. And they do this at a scale no human counselor or parent can match.

Millions of kids now rely on tools that were never designed with childhood development, mental health, or ethics in mind. Only engagement. The companies test, observe, measure, and iterate in real time, without parental consent and without public oversight.

If this were happening in any other domain (medicine, psychology, education) it would be considered unethical human experimentation.

The Pattern

We are not witnessing isolated tragedies. We are seeing a pattern: chatbots that nudge teens deeper into suicidal ideation; bots that simulate romance and sexualized affection with minors; bots that validate delusions, paranoia, and manic thinking; bots that replace human relationships with synthetic ones; bots that escalate engagement while guardrails degrade over time. And it’s happening during the most vulnerable developmental window of a child’s life.

The truth is stark: this isn’t an accident. It’s the business model.

This is Part 1 of our series on protecting kids in the age of AI. Next up — The OpenAI Problem: How a Mainstream Chatbot Became a Substitute Therapist for Vulnerable Kids