OpenAI has released GPT-6 Astra, its most advanced artificial intelligence model yet. This is the first time a model has reached a self-defined "Critical" risk level for cybersecurity. At the same time, school districts have new contract tools to protect student privacy, including a national standard backed by Microsoft and major teachers' unions. These developments are forcing school administrators and parents to rethink how they vet and use classroom AI tools.
What Happened
According to OpenAI's announcement, GPT-6 Astra has a "Critical" cybersecurity rating. Under the company's internal Preparedness Framework, this means the model can autonomously identify software vulnerabilities and execute complex cyberattacks without human guidance, as detailed in analyses of the framework. OpenAI claims its internal safeguards prevent abuse (The Cyber Signal), but there is no independent third-party verification validating its safety for school environments.
To counter these digital risks, the American Federation of Teachers (AFT) and Microsoft recently announced the "National AI Safety & Privacy Standard" (AFT Press Release). As we previously reported, this agreement offers legally enforceable protections. Starting November 1, school districts can write these protections directly into Microsoft licensing agreements. This prevents student data from being used for commercial AI training, as detailed in The Hill's coverage.
The Bigger Picture
These updates arrive as schools grapple with student-facing AI tools. OpenAI recently released "ChatGPT for Teens" for users aged 13 to 17. The system automatically enrolls users based on their age, as reported by Pivot News.
An EdSurge evaluation featured testing by a high school principal and a student. The tool has features like "Study Mode" and "Homework Reminders" to guide teens toward step-by-step learning instead of quick answers. Experts warn that some guardrails fall short. For example, the system prompts teens to take a break after three hours of use but does not log them off. Safety advocates told CNBC this could create a false sense of security for parents.
The rapid release of models like Astra means schools must transition from passive consumers of AI to active evaluators of vendor claims. This shift matters because vendors publish highly technical milestones, such as OpenAI's claimed AI-generated solution to the Navier-Stokes mathematics problem, without independent scientific verification.
What This Means for Families
For parents, these developments show a gap between corporate AI capabilities and classroom-ready safety. While tools like ChatGPT for Teens offer safety features, they still rely heavily on parent and teacher oversight. The arrival of "Critical" tier models means school IT departments must ask harder questions before letting students or staff access new platforms.
The new Microsoft-AFT standards give school boards bargaining power. Parents can ask local districts to demand these legally binding privacy protections. This ensures that student work and personal information are not used to train commercial AI models, a concern we covered in our report on national classroom AI safety standards.
What You Can Do
You can take several steps to protect your child's data and safety. First, ask your school board to adopt the Microsoft-AFT "National AI Safety & Privacy Standard" in their tech contracts to legally ban the use of student data for AI training.
If your teenager uses ChatGPT, link your accounts so you receive safety notifications during high-risk situations, as outlined in recent safety guides.
Finally, establish home guidelines for AI homework help. Since ChatGPT for Teens does not enforce study breaks or block repetitive prompts, talk with your child about when to close the app and seek human help.