Introduction

The internet has given rise to new legal battles, and a recent federal appellate ruling has sparked debate by dismissing a possession charge against a man with AI-generated child sexual abuse material. The decision hinges on long-standing First Amendment protections, even as experts warn of real harms to children.

What Happened

Steven Anderegg, a software engineer, used generative AI tools to create hyper-realistic images depicting child sexual abuse. He allegedly possessed, distributed, and even sent the material to a minor via Instagram direct messages in October 2023. Meta's safety systems detected the images and reported them to the National Center for Missing and Exploited Children's CyberTipline, prompting a law enforcement investigation and his arrest. He faced multiple charges, including knowingly possessing CSAM and distributing it. However, a district court dismissed the possession count, citing First Amendment precedent. The U.S. Court of Appeals for the 7th Circuit ultimately agreed, dropping the possession charge while leaving production and distribution counts intact.

Why This Matters

Legal scholars explain that the ruling rests on decades-old Supreme Court decisions. In Stanley v. Georgia, the Court affirmed a right to possess obscenity in the home. Later, in Osborne v. Ohio (1990), the Court carved an exception for real-child CSAM, reasoning that possession fuels demand and re-victimizes children. A 2002 ruling then extended that protection to real children only, striking down a ban on entirely computer-generated images. The 7th Circuit panel acknowledged the dilemma posed by AI's rapid advancement but said it cannot redraw constitutional lines on its own. Experts like Dr. Mary Anne Franks argue the decision normalizes the sexualization of children and undermines bodily autonomy, even if it follows existing precedent.

Key Takeaways

  • The appellate court upheld First Amendment protections for AI-generated CSAM in the home, citing prior Supreme Court rulings.
  • Steven Anderegg's possession charge was dropped, but he still faces charges for producing and distributing the material.
  • Meta's detection systems and the CyberTipline played a key role in bringing the case to light.
  • Experts warn that AI-generated CSAM normalizes child sexualization, regardless of whether a real child was depicted.
  • The government may still appeal to the Supreme Court, potentially arguing that AI models were trained on real CSAM imagery.

Conclusion

The ruling highlights a significant gap between existing constitutional law and the realities of generative AI. While the court adhered to precedent, the decision leaves many concerned about the normalization of child exploitation in digital spaces. As AI tools become more sophisticated, the legal system faces mounting pressure to update protections—without overreaching into protected speech. Stakeholders, policymakers, and advocacy groups will likely continue pushing for narrowly tailored laws that address AI-generated harms while respecting constitutional boundaries.