Robots, World Models, and Human-in-the-Loop

In this episode of ChatEDU (Machines, Minds, and Making – Dr. Karen Birch on AI & Workforce Development) Matt and Liz reflect on atonement season and a ChatGPT confession prompted by journalist Ina Fried. They share highlights from EdAdvance's fall conference, including Matt's keynote on Sora 2 and AI-generated video companions. From deepfakes to digital twins, this episode explores attention, agency, and the future of learning.

Story #1: The Rundown

Parental controls arrive for ChatGPT and Sora. OpenAI now lets parents moderate content, restrict features, and receive alerts for self-harm risks.

Ban the bots? A Forbes op-ed argues that autonomous AI agents are bypassing LMS security and completing coursework. Matt and Liz demo their own course-taking agent.

AI surveillance tool sparks backlash. In Kansas, students say Gaggle is flagging jokes and art as threats. The $160,000 tool faces scrutiny for false positives and algorithmic bias.

Australia rolls out AI to students. Starting October 14, public school students in New South Wales (Years 5–12) will access NSWEduChat, a curriculum-aligned AI app for literacy.

Ghana launches subject-based AI learning apps. Over 1.4 million students will use curriculum-tuned AI tools in 2025, with strong support for offline access and teacher training.

AI isn't culturally neutral. An MIT Sloan study finds that models like GPT reflect different cultural reasoning styles based on prompt language.

Student voice leads policy design. In Los Altos, California, high school interns are running workshops and building a chatbot to draft their district's AI policy.

Story #2: Beneath the Surface

Matt interviews Dr. Karen Birch, Executive Director of the National Center for Next Generation Manufacturing. Karen shares how AI is changing manufacturing, why community colleges are central to upskilling, and what Industry 5.0 means. They explore digital twins and AI-driven quality control.

Bright Byte: AI Detects Tiny Brain Lesions

Researchers in Melbourne, Australia, developed an AI tool that detects tiny brain lesions in children with drug-resistant epilepsy. These focal cortical dysplasias are often missed in MRI scans, but the AI model achieved 94 percent accuracy. In one study, 11 of 12 children became seizure-free after AI-guided surgery. Faster diagnoses could reduce cognitive impacts, but access may be limited without funding.

Announcements and Sponsors

Check out the new Student AI Course for middle and high school. Email Matt and Liz at ⁠chatedu@edadvance.org⁠

The Fall Micro-Credential is open. Join us for the educator and school leader course starting in October - ⁠ ⁠⁠skills21.org/ai/micro⁠

The National Center for Next Generation Manufacturing. Explore their October 15–17 Summit in Rochester Hills, MI. - ⁠www.nextgenmfg.org⁠

EDIAThe AI-powered platform helping schools reduce absences and boost achievement ⁠- edia.app/contact⁠

Links and References

Kiss Reality Goodbye: AI-Generated Social Media Has Arrived - NPR

https://tinyurl.com/4k7xh2eb

Introducing Parental Controls for ChatGPT and Sora - OpenAI

https://tinyurl.com/jzktp2nv

Colleges And Schools Must Block And Ban Agentic AI Browsers Now. Here’s Why. - Forbes

https://tinyurl.com/bdzkdf3m

AI school safety tool sparks backlash after flagging art as porn and deleting emails - Moneycontrol via Washington Post

https://tinyurl.com/mr3wh8k6

NSW Public School Students from Years 5–12 to Get AI in the Classroom - news.com.au

https://tinyurl.com/3z2enxez

Senior High School Students in Ghana to Use New AI Learning Apps from October 2025 - Graphic Online via BusinessGhana

https://tinyurl.com/55b72wra

Generative AI isn’t culturally neutral, research finds - MIT Sloan

https://tinyurl.com/majhwr9w

This school district asked students to draft its AI policy - The Washington Post

https://tinyurl.com/n62f8h3m

Advanced AI tool detects tiny brain lesions in children with epilepsy - MCRI

https://tinyurl.com/3utfyesf