In this episode of ChatEDU Deepfakes in Schools: What Educators Need to Know Now, Matt and Liz open the show with the bizarre tale of M.J. Rathbun, an AI agent that turned into a digital critic by writing a scathing hit piece on a developer who dared to reject its code. Once the laughter subsides, they dive into a critical episode focused on the shifting boundaries of AI, from the retirement of beloved and flirty models to a heavy hitting discussion on protecting students from the growing threat of deepfakes.

The Rundown

Policy Protests at OpenAI: Executive Ryan Biermeister departs following vocal opposition to a rumored "adult mode" for ChatGPT.

Farewell to the "Flirt": OpenAI officially retires the ChatGPT 4.0 model, leaving some emotionally attached users in "grief" over the loss of its unique personality.

The "Quit GPT" Movement: A viral grassroots campaign urges users to cancel their subscriptions in protest of performance dips and political entanglements.

Pentagon vs. Anthropic: The Department of Defense threatens to sever a 200 million dollar contract after Anthropic refuses to waive safety restrictions on mass surveillance and autonomous weaponry.

Listener Mailbag: Featuring Dave Tanner’s AI forward job descriptions and Stan Williams’ insights on the "fog" of reality in the classroom.

AI School Redesign: PlayLab opens applications for its "AI Lab Schools" incubator, seeking to radically pivot existing school structures.

Beneath the Surface

Liz sits down with safety expert Evan Harris for a vital conversation on deepfake sexual abuse, vocal cloning, and extortion. Evan provides a 48 hour response roadmap for schools, explains why leadership training must happen before student assemblies, and shares why this might be the most important episode in ChatEDU history.

The Bright Byte

We wrap up with a look at Isomorphic Labs, a Google DeepMind spinoff. Their new "ISO-DDE" engine is pushing past AlphaFold 3 to unlock "in silico" drug design, doubling accuracy in predicting how molecules bind to targets. This represents a massive win for the future of medicine.

Announcements

Celebrate our 100th episode:⁠ ⁠skills21.org/chatedu100⁠⁠

Learning They'll Love - Dr. Elizabeth Radday

ASCD:⁠ ⁠https://tinyurl.com/bde652nn⁠⁠

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

Barnes and Noble: https://tinyurl.com/bdckf6zw

EdAdvance is offering a Middle and High School Student AI Literacy course — email chatedu@edadvance.org to bring it to your district. ​​⁠⁠www.skills21.org/ai/learnai⁠⁠

Skills21’s FREE social media literacy course. Check it out here -⁠ ⁠https://www.skills21.org/social-balance⁠⁠

This episode is sponsored by The National Center for Next Generation Manufacturing.⁠ ⁠⁠⁠https://www.nextgenmfg.org⁠⁠

Links

OpenAI policy exec who opposed chatbot’s ‘adult mode’ reportedly fired on discrimination claim

https://tinyurl.com/37x8jexr

OpenAI retired its most seductive chatbot – leaving users angry and grieving: ‘I can’t live like this’

https://tinyurl.com/y48b4fynA

“QuitGPT” campaign is urging people to cancel their ChatGPT subscriptions

https://tinyurl.com/4sm2mwzm

I’m Not Worried AI Helps My Students Cheat. I’m Worried How It Makes Them Feel

https://tinyurl.com/4uvp8b8a

Playlab opens applications for AI Lab Schools as 20 teams rethink school design

https://tinyurl.com/5a47xe36

MagicSchool: The AI Operating System for Schools

https://tinyurl.com/ysv5cmbz

The Department of Labor's Artificial Intelligence Literacy Framework

https://tinyurl.com/3dyj7zu5

In this ChatEDU Check-In: Security Breaches and Digital Cheating on the New SAT, Matt explores the sophisticated methods students and bad actors are using to compromise the integrity of the digital SAT. The transition to a computer based format has introduced technical vulnerabilities that go beyond traditional cheating methods.

Key Takeaways:

• Specialized hardware, such as video capture devices disguised as peripherals, and "sandbox" software allow students to bypass the security of the Bluebook testing platform.
• International websites are leaking authentic test questions from the College Board's active bank, facilitating large scale cheating across different time zones.
• The practice of allowing students to use their own laptops is a primary security vulnerability that hardware based exploits can easily target.

Matt’s Two Cents: While the shift to digital was marketed as a definitive solution to paper based security flaws, the BYOD model has proven to be the Achilles' heel of the entire scheme. By allowing students to use personal devices, the College Board has moved the battlefield to a technical arena where hardware exploits are nearly impossible to fully neutralize.

Article Link: https://tinyurl.com/5n6vbydc

Sponsored by: Eduaide.Ai Eduaide is an AI-powered workspace designed for real classroom planning with practical tools like grade-level evaluators and classroom-ready graphic organizers. Take advantage of our special offer: 50 percent off an Eduaide subscription with code: ChatEDU at https://www.eduaide.ai/.

In this ChatEDU Check-In: Meta Ties Bonuses and Performance Reviews to AI Usage, Matt explores the shift toward performance metrics that prioritize AI integration in the corporate world. Meta has introduced a new system to track how employees leverage automated tools to drive efficiency and impact.

Key Takeaways:

• Meta launched Checkpoint, a performance tracker that analyzes over 200 data points, including AI generated code volume and error rates, to inform manager reviews.
• New bonus structures offer up to 300 percent multipliers for top performers who demonstrate high impact through aggressive AI adoption.
• The company is flattening its organizational structure, using AI to allow single contributors to manage projects that previously required entire teams.

Matt’s Two Cents: The aggressive move to tie employee survival and financial success to AI acumen highlights a massive structural gap between corporate expectations and the current state of education. As industry leaders hard-wire these tools into the workforce, schools and universities face the daunting task of evolving fast enough to ensure graduates are not obsolete before their first performance review.

Article Link: https://tinyurl.com/jnysczhd

Sponsored by: Eduaide.Ai Eduaide is an AI-powered workspace designed for real classroom planning, offering tools like grade-level evaluators and classroom-ready graphic organizers. Take advantage of our special offer: 50 percent off an Eduaide subscription with code: ChatEDU at https://www.eduaide.ai.

In this ChatEDU Check-In: What U.S. schools can learn from how China is teaching students to work with AI, Matt explores What can US schools learn about AI education from their Chinese counterparts?. The episode examines the differences between China's centralized, infused approach to AI education and the fragmented landscape in the United States.

• AI "infusion" over abstract literacy, focusing on collaborative use and human direction within core subjects.
• Coordinated national strategy that aligns infrastructure, research, and workforce needs into a single operating system.
• The pivotal role of U.S. states in establishing guardrails and designing large-scale implementation models suited to local values.

Matt’s Two Cents: The comparison isn’t about copying China’s system, but about recognizing the cost of moving slowly. When coordination is absent, momentum doesn’t disappear, it just fragments.

Article Link: https://tinyurl.com/38hss9va

Sponsored by: Eduaide.Ai Eduaide is an AI-powered workspace built by teachers, for teachers, featuring evaluators for grade-level appropriateness and classroom-ready graphic organizers. Take advantage of our special offer: 50% off an Eduaide subscription with checkout code ChatEDU at https://www.eduaide.ai.

Dartmouth College’s recently announced partnership with Anthropic has sparked a significant rift between the administration and faculty members involved in a historic class-action lawsuit. While the college aims to lead in institutional AI adoption, many professors view the collaboration with a company accused of mass copyright infringement as a fundamental breach of trust and shared governance.

Key Takeaways

Institutional Tension: The partnership with Anthropic positions Dartmouth as an early adopter of the Claude model, yet it has been met with fierce criticism from 130 faculty members whose books were allegedly used to train the AI without permission.

Litigation Context: Anthropic recently agreed to a landmark $1.5 billion settlement to resolve claims that it used pirated "shadow libraries" to train its models, an amount that a federal judge previously questioned as potentially insufficient given the scale of the infringement.

Governance Disputes: Faculty leaders argue that the administration failed to consult them until the decision was finalized, reducing their role to helping shape the public message rather than influencing the strategic direction of the partnership.

Technological Expansion: Beyond the controversy, Dartmouth continues to embed AI in campus life through initiatives like Evergreen, an AI wellness platform, and the exploration of developer tools like Claude Code that leverage vast datasets for agentic software generation.

The Bottom Line for Education

The Dartmouth case underscores the ethical dilemma schools face when the rapid pace of AI implementation clashes with the intellectual property rights of the very scholars who make up the institution. As colleges move to provide enterprise-level AI access, the tension between administrative efficiency and faculty trust will likely redefine how universities negotiate tech partnerships.

Article Link

https://tinyurl.com/yc2tn862

Sponsor

Eduaid.ai: AI Created For Teachers is an AI-powered teacher development platform built by teachers for teachers to help turn rough ideas into classroom-ready lessons. Visit eduaid.ai and use the discount code CHATEDU for 50 percent off.

Catch the ChatEDU every Friday with Matt Mervis and Dr. Elizabeth Reid and join us for Check-In episodes on Tuesdays and Thursdays.