AI for Education

It's another busy week in AI + Education, so here's our weekly roundup of the key highlights. We also shared these curated updates  in our Leading GenAI Adoption & Policy course to help keep school leaders informed on the latest developments - we know finding the time is often difficult. ✅OpenAI recently launched Sora 2, its new video and audio generation model, alongside a social iOS app with safety features. The app surged to become the third most popular AI tool after ChatGPT and Google Gemini this weekend. However, widespread copyright abuse immediately exposed gaps in OpenAI's responsible use framework, forcing the company to implement a series of changes. The app's viral nature suggests significant appeal among younger people. ✅Meta will begin conversations with its AI chatbot to personalize ads and content starting December 16, 2025, with no opt-out option for users. This marks a significant shift in how personal AI interactions become monetization tools—meaning casual queries could directly fuel targeted advertising for related products. ✅Teach For America's Reinvention Lab generated over 1,250 classroom tools in 2025 through hackathons, prototype shops, and Arcade AI. These programs take an interesting hands-on approach to AI literacy where educators learn by doing: prototyping, experimenting, and creating using design thinking methods. ✅ Two AI infrastructure startups launched last week: Tinker provides a managed API for fine-tuning large language models, while Periodic Labs is building AI systems that learn from physical experiments rather than internet data. Both signal a shift toward more specialized and accessible AI infrastructure and creative data sources. ✅The Budget Lab analyzed AI's impact on the U.S. labor market since ChatGPT's launch and (perhaps surprisingly) found no measurable economy-wide disruption in employment or occupational patterns thus far, emphasizing that better data is needed to fully understand AI's labor market impact. But with no comprehensive usage data currently available, researchers must rely on imperfect proxies like "exposure" metrics and data from individual AI tools, limiting the ability to detect early labor market shifts in one direction or another. What did you think of this week's news? Anything we missed? And if you’re interested in learning more about our course, check out the link in our comments.

What if students could get instant, thoughtful feedback without compromising their learning? We recently came across Anna Mills’ approach to AI feedback as a low-risk way for educational AI use. This approach tackles three key educational needs: 1. Preserving student voice: Students maintain ownership of their ideas and voice while still getting AI support to improve their thinking 2. Building AI literacy: Students learn to question, push back, and iterate rather than accept AI responses at face value 3. Avoiding integrity issues: Since students reflect on feedback, not having AI generate their content, their integrity stays intact After reading Anna's work on the topic, we had to have her on a webinar. We're excited to announce she'll be joining us for a free webinar next Tuesday, Oct. 7th. In this session, educators will learn strategies to help students intentionally engage with AI feedback—building essential skills for navigating AI in all areas of life. Key topics we will cover: • The Peer & AI Review + Reflection (PAIRR) feedback prompt: Designed by educators to help students see their work through readers' eyes—identifying strengths and areas for revision • Scaffolding critical engagement: Seven practical strategies for teaching students to iterate with AI feedback, including how to push back on suggestions, explore uncertainties, and seek contradictory perspectives. • Protecting human connection: Why teacher and peer feedback remain essential, and how AI can supplement rather than replace these interactions. • Hands-on practice: Work with real examples to test iteration strategies in action. Can’t make the time? A recording + resources will be emailed to all registrants. Links in comments to register.

🤯 STOP Teaching AI... Start Teaching AI LITERACY! The conversation is not about banning ChatGPT. It's about a fundamental shift in how we educate. We sat down with Amanda Bickerstaff, founder of AI for Education, and she revealed why her S.E.E. Framework for AI Literacy is critical for every school. Amanda's company has grown from a single laptop to supporting 250+ institutions because she focuses on the hard truths about AI in the classroom. Ready to pivot your school's strategy from fear to future-ready? Watch or Listen to the full interview now! YouTube: https://lnkd.in/gqgBFufG Apple: https://lnkd.in/gvjdzbD8 Spotify: https://lnkd.in/ggyeux2y What's your school's biggest challenge right now: safety, ethics, or effective use of AI? Let us know in the comments! #AILiteracy #EdTechLeadership #FutureofLearning #IntegrateThis #EducationPolicy #AmandaBickerstaff

It's another eventful week in AI + Education, so here's our weekly roundup of the key highlights worth staying informed about (for more details, check out the slides): ✅Mashable reviewed three foundation model learning modes —Google's Gemini Guided Learning, ChatGPT, and Anthropic's Claude—finding that each of the modes had significant drawbacks and are best suited for self-directed learners. ✅A Massachusetts teacher's review of AI-graded exams uncovered a significant error: the technology had incorrectly scored 1,400 MCAS essays across 145 districts. Though this represented a fraction of the 750,000 essays graded, the discovery highlighted the challenges of implementing new AI technology without significant human oversight. ✅The Digital Education Council analyzed 106 global case studies, identifying 24 AI methodologies for student engagement. The report finds that students increasingly seek AI feedback rather than consulting faculty and peers. This shift suggests that while AI can free up faculty for mentorship, poorly designed tools may weaken human interaction and engagement—highlighting the need for thoughtful pedagogical integration. ✅The Chan Zuckerberg Initiative’s Learning Commons is releasing their Knowledge Graph and Evaluators as training data sets for AI tools to improve pedagogical outputs. CZI worked with Anthropic’s Claude to support its AI-generated pedagogical content to be aligned with state standards and grounded in research-based learning pathways. The tools are transitioning from private beta to general availability in 2026. ✅ A recent Pew Research Center survey found that Americans are significantly more concerned than excited about AI's increasing role in daily life (50% vs 10%). Most believe AI will weaken core human capabilities like creativity and relationships, though they're open to AI’s help with routine tasks. What did you think of this week's news? Anything we missed? Links in the comments!

I’m excited to share the final session in our 3-part webinar series leading up to EDTECH WEEK 2025: We’ll hear from: - Lalitha Vasudevan – Vice Dean for Digital Innovation; Director, Media and Social Change Lab, Teachers College, Columbia University - Dr. Ellen Meier – Executive Director, Center for Technology and School Change, Teachers College, Columbia University - Amanda Bickerstaff – Founder, AI for Education They’ll share real-world insights on how AI is being used in classrooms—and what it means for students, teachers, and schools. 🎫 Save your spot here: https://lnkd.in/eUDBpiEE #AIinEducation #FutureOfLearning #EdTechWebinar Ben Kornell Ash Kaluarachchi Emily Hollender StartEd EDTECH WEEK Edtech Insiders Sarah M. Sharon Don Jen Lapaz Benny Oreo

Buying an AI product isn't an AI strategy for your school. Unfortunately, many people in my shoes want you to believe that their product will address gaps that require intentional AI literacy, professional development, policy and more. At Brisk Teaching, we know that we're just a part of a broader ecosystem of opportunities and challenges in this moment. That's why, I'm excited that our very own Elisabeth Bostwick will be in conversation with Amanda Bickerstaff from AI for Education to cover the best practices of Developing a GenAI Adoption Plan. Webinar link in the comments!

It's been another eventful week in AI + Education, so here's our weekly roundup of the key highlights worth staying informed about: ✅ Two bipartisan bills are advancing through Congress to boost AI literacy, giving a strong indication of where the government is heading on AI education policy: • The RAISE Act, while not requiring AI standards, would encourage states to develop AI academic standards for K-12 • The NSF AI Education Act takes a broader approach with scholarships, fellowships, and up to eight regional community college AI centers focused on developing career pathways and best practices. ✅ OpenAI's largest ChatGPT usage study analyzed 1.58 million conversations, revealing that three primary use cases—practical guidance, writing assistance, and information seeking—account for 77% of all interactions. Nearly half of all users are between 18 and 25 years old, highlighting the platform's strong appeal to younger demographics. ✅Following multiple lawsuits and public scrutiny regarding student safety, OpenAI is developing enhanced protective features specifically for younger users, including age prediction technology that can automatically detect underage users and redirect them into a more age-appropriate version. ✅ Anthropic's third Economic Index Report shows significant shifts in AI adoption patterns. Educational instruction tasks have surged by over 40% since December 2024, while fully automated tasks jumped from 27% to 39% over nine months. API users now automate 77% of their interactions compared to just 49% for individual consumers.  The report also reveals a digital divide: higher-income countries use AI collaboratively across diverse fields, while lower-income nations rely more on automation for coding tasks. ✅ Both OpenAI's GPT-5 and Google's Gemini achieved gold-medal performances at the 2025 ICPC World Finals programming competition, outperforming top university teams. In a first for any AI system, GPT-5 completed every challenge with a perfect score on all 12 problems. What did you think of this week's news? Anything we missed? Links in the comments!