‘This Is The Future Of Pedagogy’: How European B-Schools Are Using AI To Rethink Teaching

When students in Imperial Business School’s Global Online MBA program ask their professors for help, the answer might not come from a human being at all.

Instead, it might come from their professor’s digital twin — an AI-powered avatar trained on module content, lecture transcripts, and live session video data, and designed to deliver contextualized, real-time support in the professor’s voice and likeness.

“The students are spread all over the world,” says Sean O’Grady, lead learning designer at Imperial’s IDEA Lab, the school’s in-house edtech innovation unit. “And often they’re asking questions on the forum or by email and waiting days for a response. So we thought: Why not create an AI version of the professor that can answer those content questions 24/7?”

The IDEA Lab piloted the digital twins in the spring term across a small number of modules, with promising results. Faculty who opted in were asked to record personalized video clips and share course content so the bots could be trained on their delivery style and academic material. “We never wanted to impose this on anyone,” O’Grady says. “We worked with faculty we’d built trust with. For some, the idea of a bot impersonating them was a little unnerving. Others were really keen.”

‘THANK YOU, PROFESSOR BOT’

The results were fascinating — not just in how students used the digital twins, but in how they interacted with them.

“In our generic chatbots, students are very transactional. They just type ‘Tell me the answer.’ But when it was a professor avatar — sitting there in video form, speaking to them — they were much more conversational,” O’Grady tells Poets&Quants. “They’d say, ‘Good morning, Professor. Thank you so much.’ They treated it like a real discussion, with politeness and reflection.”

In some modules, the twin was deployed not just for generic Q&A but to support specific learning activities. In one organizational behavior course, students were asked to read a case and reflect on six or seven questions. “Instead of just writing a response, they were prompted to talk through their thinking with the twin first,” O’Grady explains. “The twin might say, ‘Well, have you considered this angle?’ Then students would revise their thinking and post their answer — with a note on whether the twin had changed their mind.”

It wasn’t always seamless. Transcripts occasionally revealed quiz answers embedded in past lecture discussions. Some students preferred more guidance, while others felt ambivalent. But O’Grady believes the potential is huge  —  especially when blended with simulation and roleplay.

In one entrepreneurship course, students negotiated investment terms with an AI-powered VC chatbot named Alex. “He had access to their business plans and red lines — like ‘Don’t accept more than 20% equity’  —  and the students had to convince him,” O’Grady says. “It’s that kind of scenario-based learning, where they actively use the concepts in real time, that I think will be transformative.”

AT VLERICK, AI GRADES THE EXAMS  —  BEFORE STUDENTS SEE THEM

While Imperial is creating digital twins, Vlerick Business School in Brussels, Belgium is deploying AI in a very different way: to assess assignments themselves for vulnerability to AI-generated cheating.

When ChatGPT went public in late 2022, says Steve Muylle, associate dean of digital learning, “everything changed.” By early 2024, Microsoft Copilot was rolled out to every student account at Vlerick  —  automatically and without warning. “Suddenly it wasn’t just optional. It was built in. And our faculty started asking: How do we make sure our assessments are still valid?”

Muylle’s team built a GPT-powered tool that allows faculty to test their assignments before releasing them to students. “You paste the assignment in, and the tool flags which parts are low, medium, or high risk of being easily completed by generative AI,” he says. “It then gives recommendations on how to revise it.”

But the goal isn’t to ban AI. Instead, Vlerick now categorizes assignments as either AI prohibited or AI encouraged. The middle ground — “light” or “moderate” use — has been phased out. “It’s too hard to police. Did they use AI just to polish the language, or to generate the content entirely? We’d rather be clear.”

RETHINKING ASSESSMENT FROM A TO Z

Vlerick’s approach is guided by Bloom’s Taxonomy, a hierarchy of learning objectives used across higher education. In a recent book chapter co-authored by Muylle (AI in Education: The Urgency of the Now), faculty are asked to map which parts of their assignments can be AI-augmented — and which must remain human.

“It’s about being intentional,” Muylle says. “If you assign a McKinsey-style strategy report, AI can help with the layout and structure. But the student still needs to bring critical thinking, synthesis, judgment. Same with a presentation — AI can help design slides, but it can’t pitch authentically. That’s human.”

Student reactions vary. “Some are eager, others avoid AI because they’re worried about plagiarism,” he says. “But we’ve seen clear differences. In classes where students rely too much on AI, they perform well on assignments — but bomb the paper-and-pencil exam.”

That insight has led to redesigned assessments, more in-class discussion, and even multi-agent simulations where students interact with AI-powered teammates. “We want them to learn how to collaborate — with machines and with humans,” Muylle says. “Because in the workplace, that’s the skill that matters.”

PORTO TAKES A SYSTEMIC APPROACH

At Porto Business School, AI isn’t an add-on or an elective. It’s embedded into the very DNA of the curriculum — at every level, for every learner.

That’s by design, says André Santana, research lecturer in data analytics and AI. “Teaching with AI and teaching about AI are not the same,” he says. “Our goal is to do both — and to do it across classroom management, faculty development, and student learning.”

AI is used to help instructors prepare teaching materials, generate synthetic datasets, and test classroom scripts. It’s also woven into courses as both a subject of instruction (machine learning, prompt engineering, ethics) and a tool of instruction. Students build proofs of concept with generative models, test operational research techniques, and compare traditional approaches to AI-driven ones.

What matters most, Santana says, is intention. “Every use of AI must align with learning goals. It’s not just about efficiency or faster responses. It’s about helping students develop critical judgment, autonomy, and the ability to choose the right technique for the right problem.”

AI ≠ EASY BUTTON

Santana is quick to note that AI can accelerate learning — but not replace it. “Having a good answer doesn’t mean you understand,” he says. “Delivering an output isn’t the same as building a competency.”

That philosophy extends to assessment as well. At Porto, human judgment is central: “Even if AI assists with feedback or preparation, the evaluation must remain human-led. Students are advised to reflect clearly on what they did themselves — and what was augmented.”

The biggest challenge, Santana says, is not technical. It’s cultural. “We need to distinguish what belongs to humans from what belongs to machines. AI can support creativity and insight — but it can’t inspire. People inspire people.”

THE FUTURE IS ALREADY HERE

Across these three leading European B-schools, AI is no longer a theoretical trend. It’s a tangible force reshaping how students learn, how professors teach — and how institutions define value.

“It’s an exciting time,” says Imperial’s Sean O’Grady. “The question isn’t whether AI belongs in the classroom. It’s how we use it — and how we preserve what makes education fundamentally human.”

DON’T MISS PORTO BUSINESS SCHOOL DEAN JOSÉ ESTEVES ON AI & THE FUTURE OF MBAs and BUSINESS MASTER’S GRADS FACE A TOUGH JOB MARKET IN THE U.S. — BUT EUROPE OFFERS HOPE

© Copyright 2025 Poets & Quants. All rights reserved. This article may not be republished, rewritten or otherwise distributed without written permission. To reprint or license this article or any content from Poets & Quants, please submit your request HERE.