Educast 3000

Hosts Ryan Lufkin and guest co-host Mary Styers are joined by Dr. Reshma Gouravajhala, Director of Impact and Research at Riipen, a work-based learning platform where educators, employers, and learners collaborate on real industry projects. Her team recently wrapped a study of more than 650 recent graduates and current students, and 74% of the graduates said a real employer project motivated them to finish their program. Gouravajhala share her thoughts about about micro internships, what employers still want in the age of AI, and why all three have a grudge against the term “soft skills.”

In this episode:

• Why work-based learning works best in small, scaffolded doses
• How a 60-hour project helped one student start her own marketing firm
• What a second-year accounting student could learn from a real client project
• Why students stay when they see the point of a course
• What small businesses get from student projects
• Social capital and who gets access to it
• Why employers still ask for durable skills
• Why a micro internship may outweigh a job fair

https://www.riipen.com/
https://www.riipen.com/project-based-internships 

Host Ryan Lufkin and co-host Jody Sailor are joined by Steven Olinger and Joe Godsey, faculty in the Business, Engineering, and Industrial Technology division at Southwest Virginia Community College. Earlier this year, the college won a nearly $500,000 SPARK grant from the Virginia Department of Energy, and this summer Olinger and Godsey ran its first five-week energy career camp for students ages 14 to 17. They talk with Ryan and Jody about fitting eight certifications into one summer, holding teenagers’ attention in the middle of summer break, and why the snack cart made it into their advice for any other colleges trying to spin up similar programs. 

In this episode:

• Olinger’s path from a family of machinists to the classroom, and Godsey’s 10 years as a design engineer
• Why the camp teaches one certification at a time
• Making safety training stick with a 15-year-old, using harnesses, ladders, and safety data sheets
• Jobs students didn’t know existed, and didn’t know they could do without leaving home
• Working lunches with local employers and the college’s advisory committees
• Credit for prior learning and what five weeks on campus means for first-generation students
• What they’d change next summer, and their advice for colleges building something similar

For further reading: 

• SPARK Energy Career Summer Camp, Southwest Virginia Community College: https://www.sw.edu/summercamps/spark-energy-career-summer-camp.html
• "Southwest Awarded Spark Grant" (college news release): https://www.sw.edu/news/2026/02/southwest-awarded-spark-grant.html

Hosts Melissa Loble and Ryan Lufkin are joined by Dr. Leslie Kennedy, Assistant Vice Chancellor for Academic Technology Services at the California State University Office of the Chancellor, where she co-leads the academic side of the CSU's system-wide generative AI initiative. Kennedy's team gave every CSU campus access to an enterprise AI environment, built the CSU AI Commons, and surveyed the system about how people are actually using AI. She talks with Melissa and Ryan about what the survey found, what it missed, and why she feels bad for her students with AI stress.

In this episode:

• Kennedy's path from linguistics to ed tech, by way of five years at WebCT
• Why portfolios stuck and micro-credentialing hasn't yet
• The equity problem that led the CSU to buy AI for everyone
• What the system-wide survey showed about student and faculty use
• The question the survey didn't capture: am I allowed to use this in this class?
• What faculty say they need, starting with assessment help
• Why administrators have a harder time admitting they're lost
• Kennedy on leaning in while the headlines call for tapping the brakes

For further reading:

• CSU AI Commons: https://genai.calstate.edu/
• Faculty AI Showcases: https://genai.calstate.edu/communities/faculty/csu-artificial-intelligence-educational-innovations-challenge-award-showcases?combine=literacy&combine_1=&field_campus_target_id=All
• CSU Systemwide AI Advisory Committee: https://genai.calstate.edu/about/systemwide-generative-ai-advisory-committee
• CSU AI Workforce Acceleration Board: https://genai.calstate.edu/about/workforce-acceleration-board

Sean Murphy, director of workforce and post-secondary pathways at the Bipartisan Policy Center, joins hosts Ryan Lufkin and Tracy Weeks for a conversation about why the handoff between education and employment still breaks down, and what it would take to fix it. Sean has worked the same problem from four seats: a local workforce board, staff jobs for a US senator and a governor, corporate philanthropy at Walmart, and now federal policy. Together they get into the four parts of a talent system that actually functions, why data built for compliance fails the people it's meant to serve, what digital credentials still leave out, and what institutions can start on now instead of waiting on legislation.

In this episode:

• Why silos are normal and bridges, not demolition, are the goal.
• The four moving parts of a working system: demand, recognition, gaps, and matching.
• How data built for compliance fails the workers and employers it's supposed to serve.
• Why most digital badges carry only a school name, a program name, and a date.
• Sean's own transcript: three remarkable quarters, recorded as 15 credits each.
• What skills-based hiring got wrong, and why a degree stays legally defensible.
• Skills as a language of learning, employment, and service, not a degree replacement.
• The people the system forgets: everyone between school and unemployment.
• One thing a registrar or CTE program can start on this semester.

For further reading:

• https://bipartisanpolicy.org/hub/ai-and-workforce-navigator/
• https://bipartisanpolicy.org/report/a-nation-at-risk-to-a-nation-at-work-the-case-for-a-national-talent-strategy/

Most teachers will never meet a psychometrician, but every standardized test their students sit for went through one. Russell Ligon and Alexandra Lee do that work,, and in this replay hosts Ryan Lufkin and Melissa Loble ask them how they landed in a field almost nobody plans on, how assessment practice has shifted over their careers, and why the technical work behind a test decides whether it's fair to the student taking it.

In this episode:

• Failure gets treated as the thing assessment exists to prevent, and the group argues it teaches more than a clean score ever does.
• Accountability has moved from centralized to decentralized models, and with that came the expectation that a test be built against the standards of the state that adopted it rather than shipped intact from somewhere else.
• Psychometricians are the reason a test can be called equitable, and Russell and Alexandra walk through the parts of the job that never show up on a score report, including how each of them found their way into it.
• Assessment of learning, for learning, and as learning are three separate jobs, and a balanced system runs all three instead of asking one test to carry the whole picture.
• Every assessment spends time that instruction could have used, and the group is direct about that trade instead of treating more student data as free.
• Validity holds the whole structure up, because a test without it still measures something, just not what anyone needed to know, and a teacher who doesn't trust the instrument won't act on what it tells them.
• What comes next in assessment turns less on new technology than on whether the culture around testing lets a test show real growth, and the conversation closes on where AI fits into that.

The loudest voices on AI in education sit at the edges, and almost nobody who teaches every day lives out there. Matt Miller calls the place they actually live the sensible middle, and this episode makes the case for it. Hosts Ryan Lufkin and guest co-host Jody Sailor talk with the creator of Ditch That Textbook, a high school Spanish teacher of more than twenty years and the author of seven books, about what changed when he stopped assigning workbook pages, why he went back to the classroom after eight years away, and what he thinks has to stay human while the ground shifts.

In this episode:

Ditching the textbook was never about hating textbooks, and Matt describes replacing the questions at the end of the chapter one at a time until he needed the book so rarely he stopped reaching for it.
The promise of AI is easiest to see in a Spanish 1 class of 31 students, where Matt can deliver strong whole-group instruction and still not get meaningful personalized feedback to every kid.
Cognitive offloading worries Matt as much as it worries families, and he recounts a district parent survey that came back saying, in a variety of ways, that they're afraid AI will make their kids stupid.
Time is the barrier teachers name out loud, though Matt points to comfort as the one nobody mentions.
Asking students what they think is the move that requires no permission from anyone, and Matt walks through the four-step conversation from AI Literacy in Any Class that gets honest answers instead of self-preservation.
The one-room schoolhouse kept the entire curriculum inside one teacher's head, and Matt traces how every innovation since has pulled learning away from that center, which is why he argues the question worth asking now is what only a human brings.

ditchthattextbook.com
Matt's free newsletter, Mondays and Thursdays: ditch.link/join
AI Literacy in Any Class by Matt Miller

Learning doesn't end at graduation, and the systems around us shouldn't either. This clips episode takes that idea and follows it across a whole life. Hosts Ryan Lufkin and Melissa Loble revisit five conversations, from a charity teaching school kids to code in Australia to a summer-learning advocate to a university leader rethinking credentials, and line them up as one continuum: the skills that matter at every stage, and the gaps that open when we treat each stage as separate.

In this episode:

• Fear, not ability, is what keeps kids out of digital skills, which is why Megan Woff's Code for Schools starts with curiosity as it works toward reaching a million students by 2027.
• Summer learning loss hits low-income students hardest, and Christi Kobus Rokicki explains how a gap that opens in childhood compounds every year that learning stops.
• Durable skills like resilience and leadership can't be studied for a test, since Emily Foote argues they're built by doing real work, which matters more as AI absorbs the entry-level tasks graduates used to cut their teeth on.
• The critical thinking and communication skills the liberal arts have always taught, Joe Potvin makes the case, outlast any single job or technology.
• Credentials should recognize the learning people already carry, and Martin Bean lays out how a connected learning system would follow a person across their life instead of ending at a diploma.

Guests:

• Emily Foote, Chief Growth and Strategy Officer, Saxbys 
• Megan Woff, CEO, Code for Schools 
• Christi Kobus Rokicki, Vice President for External Affairs, National Summer Learning Association
• Joe Potvin, History Adjunct Professor, Simmons University, and Senior Portfolio Manager, Cengage 
• Professor Martin Bean
• Katie Novak, Founder and Chief Education Officer (CEO), Novak Educational Consulting

Dale Leszczynski leads RMIT University's Generative AI Lab for Education (GAILE), where he's spent the past year building a skills continuum (instead of a single checklist) for roughly 5,000 educators , on the theory that nobody stays an expert in this field for long. Hosts Ryan Lufkin and Melissa Loble talk with Dale about what happened when other Australian university leaders mapped their own institutions against that continuum, and about the debate RMIT staged at its first education conference: are AI tools morally neutral, or does responsibility for them rest entirely on the people using them? And now that AI can do almost anything, what should humans still do themselves?

In this episode:

• RMIT's AI Skills Continuum treats AI skill-building as a moving target, not a fixed benchmark, since what counts as expert practice keeps changing faster than anyone can master it.
• When Victorian university leaders mapped their own institutions against that continuum, most landed at the earliest stage, a sign of a confidence gap more than a competence gap.
• Efficiency is the wrong goal for AI in education, since it's made his own work more expansive, not faster, which means the work itself has to change instead of just speeding up.
• Students at RMIT's new AI advisory board raised concerns about their own learning unprompted, and the thing they asked for most, across every discipline, was consistency, not more rules.

For further reading:
rmit.edu.au/gaile
https://aiskillscontinuum.com/
https://www.anthropic.com/research/global-workspace

Ryan Lufkin and Melissa Loble step back this episode and hand the conversation to two technologists who interview each other. Mike Mast, Principal Group Program Manager for Microsoft Education and a 1EdTech board member, and Zach Pendleton, Chief Architect at Instructure, trade questions about what interoperability means once AI agents start acting inside the tools educators and students already use.

The starting point: when the client is an agent rather than a person, agents can learn new vocabularies and reason over different systems on their own, so the hard part shifts from rigid data schemas to giving agents a real semantic understanding of a course. Mast points to Canvas by Instructure and its Smart Search beta as one of the few early examples of that idea in practice. Pendleton makes the case that none of this replaces the open standards ed tech already runs on. He frames MCP, LTI, OneRoster, and Caliper as an "and," not an "or."
From there the two get into the part that keeps Pendleton up at night: as actions move further from the human who set the goal, how do we keep people in authority? They cover delegated authorization and consent, audit trails, the gap between what AI can do and what we want it to do, and who carries the cost when an automated tool gets it wrong. 

In this episode: 

• Agentic AI changes what interoperability optimizes for, moving the priority toward semantic understanding of course content rather than identical schemas across systems.
• New protocols build on existing standards instead of replacing them, so prior investments in open APIs and LTI still matter.
• Keeping a human in control sometimes costs a little speed, and both guests argue that maximum speed should not be the goal of an educational AI system.
• Transparency is non-negotiable: if AI touches a grade, students should know, and they should be able to question the result.

James Genone has spent his career at the meeting point of two fields most people keep apart: philosophy and cognitive science on one side, education technology on the other. In this episode, Northeastern University's Senior Vice Chancellor for Learning Strategy joins Ryan Lufkin and Melissa Loble to argue that the design behind a tool decides whether AI deepens learning or just does the thinking for students.

His test is simple: does the AI behave the way a good human tutor would?

Drawing on his founding work at Minerva University, his product role at Atypical AI, and his current advising with the Digital Education Council, Collage AI, and SkillBench, James lays out what learning science actually asks of these tools and where institutions keep going wrong.

In this episode:

• The tutor test James applies to any AI assistant: a good tutor never writes the student's essay, and the AI shouldn't either
• Why engagement is the real engine of learning, and how connecting a subject to what a student cares about drives the attention that learning depends on
• His case against grades, built on an investor analogy: you can't tell two companies apart if all you see is a letter grade, and a transcript does the same thing to students
• The two ways institutions keep getting AI adoption wrong, either leaving faculty to fend for themselves or handing down blanket mandates, and the scaffolded, faculty-led approach he'd put in their place