AI-TELL Academy: How Penn State Trains Engineering Faculty to Teach with AI (Without Creating ‘Artificial Understanding’)

The Problem Every Engineering Educator Is Quietly Facing

AI tools are in the classroom whether faculty invite them or not. Students are using ChatGPT to draft lab reports, debug code, and work through thermodynamics problems. The question isn’t if AI is present — it’s whether educators have a coherent strategy for it.

Most don’t. Not yet.

That’s the gap Penn State’s Leonhard Center for the Enhancement of Engineering Education set out to close with the AI-TELL Academy — Artificial Intelligence for Teaching Excellence and Lifelong Learning.

What the AI-TELL Academy Actually Did

From June 2–4, a cohort of College of Engineering faculty gathered for a structured, hands-on three-day workshop. This wasn’t a passive sit-and-listen event. The programming was built around real faculty questions — gathered through pre-event surveys and one-on-one consultations — and designed to produce tangible outputs faculty could bring directly into their classrooms.

Sessions covered:

• Prompt engineering — how to construct effective AI prompts and teach students to do the same
• Human-centered AI integration — keeping students, not tools, at the center of learning
• AI-resilient and AI-integrated assignment design — building assessments that work with or without AI, depending on the learning goal
• AI syllabus policy development — giving faculty language and frameworks to set clear expectations

The organizers — Ibukun Osunbunmi, Stephanie Cutler, Sarah Zappe, and Bono Shih — didn’t guess at what faculty needed. They listened first, then built.

The Core Insight: Don’t Let AI Create Artificial Understanding

This is the line that cuts through all the noise.

Ibukun Osunbunmi, assessment and instructional support specialist and assistant research professor at the Leonhard Center, put it plainly: “A student who cannot do an exercise, solve a problem or do an engineering task without artificial intelligence will not be able to evaluate if an AI outcome is good or not.”

That’s not an anti-AI argument. It’s a pro-competence argument.

If a student can’t solve a structural load problem from first principles, they have no baseline to judge whether the AI’s answer is reasonable, flawed, or dangerously wrong. In engineering — where outputs have real-world consequences — that gap isn’t academic. It’s a safety issue.

The AI-TELL framework pushes faculty to make deliberate choices: when should AI be integrated into a task, and when should it be excluded to protect foundational skill-building? Both are valid answers. Neither should be accidental.

Why Three Days? Because Depth Requires Time

Stephanie Cutler, director of assessment and instructional support, was direct about the format decision: “A brief workshop wouldn’t be sufficient. We wanted to provide faculty a chance to dive in, create class materials, gain experience working with AI tools and walk away with things they could use in their classes.”

This is a model worth noting for any organization running AI adoption programs.

Short workshops create awareness. Immersive workshops create capability. If you want faculty — or employees, or teams — to actually change behavior, they need time to experiment, fail, ask questions, and iterate. Three days of structured hands-on work does what a 60-minute webinar simply cannot.

The AI Tools Dimension: What Faculty Actually Worked With

The academy gave faculty direct, hands-on experience with AI tools in an educational context. While the program wasn’t a product review session, the practical focus on prompt engineering and assignment design means faculty were working through the real mechanics of tools like ChatGPT and similar large language models.

For educators looking to replicate this approach, the relevant tool categories include:

What Faculty Took Away

Post-event feedback captured the shift in confidence the program created. Jared Butler, assistant professor in the School of Engineering Design and Innovation, wrote: “Confronting the challenges associated with AI in academia can be daunting. While most of us are still unsure of what the future holds in this sphere, the next steps feel steady and promising.”

That’s the realistic win here. Not certainty — nobody has that. But a clearer path forward and practical tools to walk it.

The Lifelong Learning Layer

The “LL” in AI-TELL isn’t decorative. Osunbunmi was explicit: “AI is not what it was last year. And what AI is today is not what it will be next year.”

This is the part most AI training programs miss. They treat AI literacy as a one-time certification rather than an ongoing practice. The AI-TELL Academy framed the workshop as an entry point into continuous learning — not a finish line.

For faculty, that means staying connected to resources, revisiting policies each semester, and treating AI integration as an evolving design problem rather than a solved one.

How to Apply This Framework Outside Academia

The AI-TELL model isn’t just for engineering professors. The underlying logic applies to any organization trying to integrate AI tools without hollowing out the human expertise that makes those tools useful.

Three principles worth borrowing:

1. Survey before you build. Find out what your team actually struggles with before designing any AI training. The Leonhard Center did this — and it made their programming sharper.
2. Design for both integration and exclusion. Not every task should involve AI. Decide deliberately which workflows benefit from AI augmentation and which require unassisted human judgment to build real competence.
3. Invest in depth over breadth. A three-day immersive beats five one-hour sessions every time when the goal is behavior change, not awareness.

The Leonhard Center’s Ongoing Resources

Penn State’s Leonhard Center has made their compiled resources publicly accessible — including video modules on acceptable AI use, AI syllabus policy guidance, and AI-integrated teaching demonstrations. Faculty and educators outside Penn State can reach out through the Leonhard Center’s website to learn more.

That kind of open-resource approach accelerates the field. It means the work done in one cohort’s three-day workshop can ripple outward to classrooms far beyond University Park.

Conclusion

The real takeaway from AI-TELL isn’t a tool recommendation or a policy template. It’s a mindset: good teaching is still good teaching. AI changes the instruments, not the goal. And the goal — building students who can think, evaluate, and solve problems independently — matters more now than it ever did before.