Prompt No More: What Claude Plug-ins Signal for AI in Education | Ep. 114

In this episode of ChatEDU, Prompt No More: What Claude Plug-ins Signal for AI in Education, Matt and Liz open with a major Meta security flaw: hackers simply asked Meta AI to grant access to high-profile Instagram accounts, including Barack Obama's. With support entirely AI-run, affected users couldn't reach a human.

The Rundown

ChatGPT's new memory system uses "dreaming" to auto-curate user preferences over years.

Google's new AI feature for Enterprise and Education plans finds scattered files and suggests folders for them.

In a Stanford blind study, law professors preferred AI answers to student legal questions over peer-written ones 75% of the time.

In a poll of 545 K-12 educators, most said AI will outweigh the internet's educational impact; about half report no district guidance.

A new six-item tool in the ECNU Review of Education measures how much students outsource thinking to generative AI.

Bernie Sanders proposed a one-time 50% stock tax on major AI companies to fund dividends for Americans.

Jack Clark and The Anthropic Institute warned frontier models are nearing recursive self-improvement, urging international verification.

Penn GSE launched a course on how K-12 students can design, critique, and build their own machine learning systems.

Maryland's new law requires public schools to appoint AI coordinators and add AI literacy standards by June 2027.

IBM's global software competition gives university students hands-on access to professional tech tools via its AI partner.

Beneath the Surface

Matt and Liz analyze Anthropic's "Claude for Legal" expansion (12 plug-ins, 20 MCP connectors), then turn to education: how similar integrations could streamline IEP compliance, audit curriculum drift in LMSs, and run automated student risk signals.

The Bright Byte

Using a new Health and AI Policy Index on 240 healthcare AI policies, Mount Sinai investigators see an opening for health systems and academic medical centers to lead in standardizing clinical AI deployment.

Announcements

Purchase Learning They'll Love - Dr. Elizabeth Radday

Amazon - https://tinyurl.com/22t9hz77

Seeing through the Machine – Teaching Bias as an Equity Practice in AI - Dr. Elizabeth Radday - https://tinyurl.com/4ybwkn7a

Check out our middle / high school Student AI Literacy course - ​​⁠⁠⁠www.skills21.org/ai/learnai⁠⁠⁠

Explore Skills21’s FREE social media literacy curriculum - ⁠⁠⁠https://www.skills21.org/social-balance⁠⁠⁠

Address Screen Time concerns - ⁠skills21.org/ai/screenshift⁠

Sponsors

The National Center for Next Generation Manufacturing: Your Gateway to Next Generation Advanced Manufacturing - https://www.nextgenmfg.org

Links

AI Policy Pressure Test

https://tinyurl.com/5xam65xs

AI Policy Scenarios Test

https://tinyurl.com/463p8err

Hackers Asked Meta AI for Instagram Access. It Worked.

https://tinyurl.com/4456m8yt

Better memory for ChatGPT

https://tinyurl.com/uz7abp49

This New Google Drive Feature Organizes Your Files for You

https://tinyurl.com/ewvdfzhv

AI Outperforms Law Professors in Stanford Law Study https://tinyurl.com/58s8vt4z

Teachers Say AI's Educational Impact Will Eclipse the Internet

https://tinyurl.com/4rh4vzye

A New Scale for Measuring AI-Driven Metacognitive Offloading

https://tinyurl.com/5f9u2428

Bernie Sanders: A.I. Is a Public Resource. You Should Own Half of It.

https://tinyurl.com/nujnbh4v

Anthropic Urges Global AI Pause Over 'Self-Improvement' Risk

https://tinyurl.com/mry54xvz

AI education for kids

https://tinyurl.com/3fzsddjt

Maryland's School A.I. Guidance Law Is Now in Effect

https://tinyurl.com/yc4xzxk6

IBM Launches Global AI Builders Challenge for Students

https://tinyurl.com/2ss23mna

Anthropic Unveils 'Claude for Legal'

https://tinyurl.com/rcf7tak4

A First-of-Its-Kind Index for Health Care AI Policy

https://tinyurl.com/49uzu77m