Problem: Generative AI is taking every industry by storm. Generally if you talk to experts in the field, they’ll agree that the future of the industry using LLMs will likely transform into many smaller models rather than one big model. Meaning, instead of having ChatGPT or Claude or Bard, people will opt for “MathGPT” or “EnglishGPT” or “AnesthesiologyGPT” where the LLM has been trained specifically on your dataset.
Ethan Mollick (a Wharton professor) briefly a Tweet about how AI could be used for teaching. This post explores that Tweet and his general framework.
Subscribe here to get access to the first 500 ideas from our blog. For just one coffee a month, you can have access to more than $500 billion dollars of ideas. What's not to love?
Solution: I write these solutions specifically in the context of Harvard Business School. I am currently a researcher at their Generative AI Lab and see AI being used across the HBS institution in 4 ways:
Non-HBS Case Discussion Augmentation (Mentorship/Simulator/Teaching): allowing students to have vocal "case discussions" with each other and with a professor in the same way that this happens in the physical classroom. Think of all the students and professors who purchase HBS cases but don't actually teach them properly because they have not been exposed to the method or have not mastered teaching or learning in that style. AI tools could potentially help here.
Teaching Note Creation (Tool/Creation): assisting professors in their ability to create new cases, but more importantly in their ability to create the teaching notes which accompany these cases (often just as tedious, but much easier to write for an algorithm since they are based off of the case themselves).
HBX Core Learning Augmentation (Teammate/Coach): especially for complex subjects, these AI tools could be used as a tutor to guide practice and learning. For instance, in HBS CORE students learn accounting. However, every student learns and makes errors differently. Based on the habits of how a student clicks on a screen or makes mistakes, the AI could provide personalized tutoring recommendations.
HBS Core Specialized Grading (Tutor): using AI as a tool for providing true evaluation and teaching for individuals who participate in online learning programs such as HBS CORE. I've participated in CORE and one of the key experiences I had was writing assignments and reviewing the writing assignments of others (today, the primary means of getting and giving student feedback is "data at scale"; however, with AI this can certainly be augmented or automated to get real-time generative feedback based on models of good work.
Contributed by: Michael Bervell (Billion Dollar Startup Ideas)