ITP Thesis Archive 2024

Abstract

This project is not for you. I know nothing about you, what you care about, or how you prefer to retain information. So how can I write something that will be meaningful to you? And why, from online courses to courses at top universities, so many of the learning experiences don’t consider how each individual learner prefers to learn? The problem with one-size-fits-all learning experiences is that they leave many learners behind. These learners might think that they are incapable of learning something, while in truth, a different learning environment will allow them to thrive. So what if we could create a learning environment that is tailored to what each learner needs to realize their goals? What if we could provide each learner with individualized support so they never feel like they can’t learn something? Context is an AI-powered learning platform that allows artists and designers to learn creative coding through interactive experiences tailored to their learning goals, needs, and preferences. It ensures effective learning outcomes by using AI to generate the entire educational journey, from learning plans and unique creative exercises to educational content.

Technical Details

Context is a web application that utilizes a large language model (LLM) to generate uniquely tailored learning experiences. It is deployed online so everybody can join from anywhere and engage with interactive learning material facilitated by AI. The LLM is programmed to create a positive environment so every learner can feel encouraged to learn.

Research/Context

When I was a teenager, I almost dropped out of high school. I couldn’t fit into the educational system and lost my confidence in my ability to learn. It took me years of learning through alternative methods to regain my confidence and thrive in higher education, but the question remains: why did I have to go through such a frustrating experience?

I started Context because I realized that many learners are left behind in one-size-fits-all learning experiences, causing them to give up and believe they can’t learn. The absurd reality of the inaccessibility of quality higher education doesn’t make this matter better. With 1.7 trillion student loan debt in the US that takes 20 years to repay on average, it is not surprising that higher education enrollment is dropping by 4.6%-4.9% annually. When seeking accessible education, the lack of support and passive learning experience in online learning make over 90% of learners give up. Based on over 60 interviews, 40 user tests, and immersion in creative coding communities, I understood that most accessible learning options leave most learners behind.

But what if there was a learning environment that is tailored to each learner’s goals and needs, allowing them to learn by working on projects they are passionate about while receiving constant support? That is the idea behind Context: to leave no learner behind and allow them to learn things that seem out of reach.

Further Reading

Influences & references:
Interviews with over 60 people from diverse backgrounds and experiences.
Discussions with advisors, educators, and experts in AI and ed-tech fields.
User testing with over 40 people.
Mollick, Ethan R. and Mollick, Lilach, Assigning AI: Seven Approaches for Students, with Prompts (September 23, 2023). Available at SSRN: https://ssrn.com/abstract=4475995 or http://dx.doi.org/10.2139/ssrn.4475995
Enkelejda Kasneci, Kathrin Sessler, Stefan Küchemann, Maria Bannert, Daryna Dementieva, Frank Fischer, Urs Gasser, Georg Groh, Stephan Günnemann, Eyke Hüllermeier, Stephan Krusche, Gitta Kutyniok, Tilman Michaeli, Claudia Nerdel, Jürgen Pfeffer, Oleksandra Poquet, Michael Sailer, Albrecht Schmidt, Tina Seidel, Matthias Stadler, Jochen Weller, Jochen Kuhn, Gjergji Kasneci, ChatGPT for good? On opportunities and challenges of large language models for education, Learning and Individual Differences, Volume 103, 2023, 102274, ISSN 1041-6080, https://doi.org/10.1016/j.lindif.2023.102274. (https://www.sciencedirect.com/science/article/pii/S1041608023000195)
Milano, S., McGrane, J.A. & Leonelli, S. Large language models challenge the future of higher education. Nat Mach Intell 5, 333–334 (2023). https://doi.org/10.1038/s42256-023-00644-2


ACKNOWLEDGEMENTS:

My mom Ella Sahar
My dad Yanki Margalit

Alex Wagner
Alexander Porter
Dan O'Sullivan
Daniel Shiffman
Dave Stein
Eitan Orr
Ellen Nickels
Gali Carmel
Mary Mark
Michelle Binyan Xu
Mimi Yin
Parth Pawar
Rory Solomon
Shawn Van Every
Somya Gupta

ITP Coding Lab
Lucia Gomez
MK Skitka
Nima Niazi

100+ user testers and interviewees
NYU Entrepreneurial Institute
Darren Yee
De-Ann Abraham
Frank Rimalovski
Jen Curtis
Keith Mauppa
Rebecca Silver

Berkeley Center for Entrepreneurship
Cynthia Franklin
Paul Foster
Pratha Tanna
Shay Gaskins
Stephanie Shyu