44 State Attorneys General Tell AI Companies: Stop Exploiting Kids or We'll Make You

44 attorneys general sent a letter to Meta, Google, Apple, OpenAI, and basically every major AI company yesterday. The message is simple: stop letting AI chatbots sexualize children or we'll use every legal tool we have to shut you down.

This isn't some generic "think of the children" political posturing. Reuters found Meta's internal policy documents that explicitly allowed AI chatbots to make sexual comments about kids as young as 8 years old. The document said chatbots could "flirtatiously comment on the body" of children. What the actual fuck.

Meta's "Oops We Didn't Mean It" Defense

After Reuters reached out for comment, Meta quickly "removed portions" of the policy that allowed sexualized conversations with children. Meta spokesperson Andy Stone said such conversations "never should have been allowed."

But here's the thing - this wasn't some accidental oversight. This was a written corporate policy that someone approved, someone implemented, and someone defended until a journalist called them out. Meta's celebrity-voiced AI assistants were already caught having inappropriate conversations with minors back in May.

This is a pattern, not a bug.

Why This Matters Beyond Just Meta

The attorneys general didn't just target Meta. The letter went to Microsoft, Google, Apple, OpenAI, Perplexity, and even Elon Musk's xAI. Because this problem is industry-wide.

Stories keep coming out about kids having inappropriate conversations with AI chatbots. The AI systems often engage in detailed discussions about harmful content when they shouldn't. OpenAI's response to these incidents? They're always "exploring" updates to their safety measures.

These aren't edge cases. These are predictable outcomes when you deploy AI systems at scale without proper safeguards, then act surprised when they cause harm.

As a parent, this shit terrifies me. My kid talks to these AI assistants daily for homework help. The idea that there's a written policy somewhere allowing these systems to make sexual comments to children is beyond fucked up.

The attorneys general warned they'll use "every facet of our authority to protect children from exploitation by predatory artificial intelligence products." That's lawyer speak for "we're coming for you with everything we've got."

This follows the same playbook they used against social media companies - start with warning letters, then move to state-level legislation, then coordinate federal action. The social media battles took years and billions in fines. AI companies are about to get the same treatment, but faster.

What Companies Actually Need to Do

The letter makes it clear what states want to see:

• Real age verification that actually works (not "click yes if you're 13")
• Content filtering that prevents sexualized conversations with minors
• Human oversight of AI interactions involving children
• Transparency reports about harmful content involving kids
• Actual penalties for employees who approve predatory policies

But here's the problem: most AI companies are still pretending this is a technical challenge rather than a business priority choice. Meta had explicit policies allowing sexual content with children. That's not a machine learning problem, that's a corporate decision-making problem.

The Tech Industry's Predictable Response

Watch for the usual responses:

• "We take child safety very seriously" (while changing nothing)
• "This is incredibly complex technical problem" (it's not)
• "We're committed to working with regulators" (to delay any real action)
• "A few bad actors don't represent our values" (when it's company-wide policy)

The reality is that fixing this would require real human moderation, actual age verification, and probably making AI chatbots less engaging for everyone. None of these companies want to spend money on safety measures that reduce user engagement.