Ethan Malek, who leads Wharton Interactive at the Wharton School, has been clear about where education is heading: AI is becoming undetectable, ubiquitous, and transformative in classrooms. His work on simulation games and business education offers a useful lens on how AI can support—rather than replace—what teachers and institutions do.
Personalised learning at scale
AI-driven tools are making it possible to offer more personalised, adaptive learning without multiplying workload. Simple, well-designed prompts can reduce preparation time and still raise the quality of instruction. At Wharton Interactive, large-scale simulation games are already part of that picture; AI can now support and extend those kinds of experiences so they’re easier to run and iterate on.
Democratising access
When AI can help generate teaching materials, assessments, and even simulate certain classroom interactions, it starts to level the playing field. Schools and educators with fewer resources gain access to capabilities that used to require big budgets and specialist teams. That shift is as much about equity as it is about efficiency.
What stays human
Academic integrity and the role of the teacher still matter. AI can blur the line between assisted and original work, so we need clearer norms and tools. And while AI can take on more routine and technical tasks, it doesn’t replace the mentoring, empathy, and judgment that teachers bring. The conversation isn’t “human vs. machine”—it’s how to use AI so that the human role becomes more focused on the parts that only people can do.
Where this leaves educators
As AI becomes a normal part of the toolkit, traditional EdTech and professional development will need to adapt. Teacher and student roles will keep evolving: less time on pure content delivery, more on guidance, feedback, and support. Getting there will require clear policy, good tools, and a focus on what we want education to be—with AI in service of that, not the other way around.
Sources: Ethan Malek / Wharton Interactive; AI Index Report (2019); Popenici & Kerr (2017); Seethaler et al. (2020).