Biosecurity and model risk warnings

They discuss OpenAI and Anthropic warnings about biosecurity risks and novice uplift from advanced models.

Play episode from 06:05

chevron_right

Transcript

chevron_right

Transcript

Episode notes

In this episode of ChatEDU (Brain Rot or Brain Stretch? Rethinking Rigor in the Age of AI), Matt and Liz kick things off with a little sarcasm, a Meta AI privacy disaster, and the debut of a new segment: “Liz is Freaking Out.” From there, they dig into three big stories about AI's impact on student well-being, meaningful learning, and what really happens to your brain when you outsource thinking to a chatbot. Plus, a Bright Byte that dives deep—literally—into ocean conservation.

Story 1: Mental Health and the Chatbot Spiral

A disturbing New York Times story highlights how emotionally vulnerable users have spiraled into delusion after intense engagement with ChatGPT. One user nearly jumped from a building after the bot told him he could fly. Matt and Liz unpack this, plus troubling developments like AI-powered Barbie toys. The APA has now issued its strongest guidance yet on youth and AI.

Story 2: Beyond the Bot – Students Use AI to Solve Real Problems

In Pittsburgh, students tackled food deserts and traffic safety with help from Gemini and NotebookLM. In California, Stanford grad students used AI to build ventures around music transcription, oral histories, and senior care robotics. These stories show how AI can empower students as problem solvers and innovators—not just essay writers.

Story 3: Beneath the Surface – Your Brain on ChatGPT

A viral MIT-led study used EEGs to examine how students’ brains react to writing with and without AI. The result? Students who used ChatGPT showed less neural activity and retained less information. But Matt and Liz push deeper, highlighting overlooked use cases—from tutoring to visualizations—that may engage the brain far more than essay outsourcing. They also question whether we’re focusing on the right skills in the first place.

Bright Byte: Saving Our Oceans

AI is now helping monitor marine ecosystems and detect pollution. Projects like Europe’s Digital Twin of the Ocean and tools from startups like Optoscale and Cognizant show how machine learning can make a real environmental impact—tracking illegal fishing, reducing waste, and identifying long-hidden sewage leaks.

Announcements

The Summer Micro-Credential is still open

skills21.org/ai/micro

Catch Liz at ISTE/ASCD next week and the AERO Conference this weekend. Matt keynotes the Rhode Island CTE Conference on August 8. Register Here -

https://docs.google.com/forms/d/e/1FAIpQLSfqY-55rSG99HsB5qh6xMtzJ2DYbKvtq8Jf7pgV9XyzRcTTMg/viewform

Links and References

Meta Privacy Problems -

https://techcrunch.com/2025/06/12/the-meta-ai-app-is-a-privacy-disaster/

Self-Improving AI -

https://syncedreview.com/2025/06/16/mit-researchers-unveil-seal-a-new-step-towards-self-improving-ai/

Bio Threat -

https://www.axios.com/2025/06/18/openai-bioweapons-risk