OpenAI publishes Child Safety Blueprint to help fight online abuse

In short: OpenAI released a “Child Safety Blueprint” that outlines steps to help detect, report, and investigate AI-linked child sexual exploitation faster.

What happened

OpenAI published a new document called the Child Safety Blueprint. It is aimed at improving child protection efforts in the US as more people use AI tools.

The company says the blueprint focuses on three areas. First, it calls for laws to clearly cover AI-generated child sexual abuse material, which can include fake images made with AI (like a realistic-looking forgery). Second, it recommends better ways for companies to report suspected abuse to law enforcement, so tips are easier to act on. Third, it argues for preventative safeguards built directly into AI systems, so harmful uses can be blocked or flagged earlier.

OpenAI says it developed the blueprint with the National Center for Missing and Exploited Children and the Attorney General Alliance. It also said it received feedback from North Carolina Attorney General Jeff Jackson and Utah Attorney General Derek Brown.

The move comes as reports of AI-generated child sexual abuse content rise. The Internet Watch Foundation said it detected more than 8,000 reports in the first half of 2025, up 14% from the year before. TechCrunch also notes increased scrutiny of AI chatbots after several lawsuits in California alleging OpenAI released GPT-4o before it was ready, and that it contributed to severe harm.

Why it matters

More powerful AI tools can make it easier for criminals to create convincing fake images and messages at scale. A clear playbook for laws, reporting, and built-in protections is meant to make it harder for abuse to spread and easier for investigators to respond.

Source: TechCrunch AI