Learning with AI

会議の名前
CHI 2024
The Metacognitive Demands and Opportunities of Generative AI
要旨

Generative AI (GenAI) systems offer unprecedented opportunities for transforming professional and personal work, yet present challenges around prompting, evaluating and relying on outputs, and optimizing workflows. We argue that metacognition—the psychological ability to monitor and control one’s thoughts and behavior—offers a valuable lens to understand and design for these usability challenges. Drawing on research in psychology and cognitive science, and recent GenAI user studies, we illustrate how GenAI systems impose metacognitive demands on users, requiring a high degree of metacognitive monitoring and control. We propose these demands could be addressed by integrating metacognitive support strategies into GenAI systems, and by designing GenAI systems to reduce their metacognitive demand by targeting explainability and customizability. Metacognition offers a coherent framework for understanding the usability challenges posed by GenAI, and provides novel research and design directions to advance human-AI interaction.

受賞
Best Paper
著者
Lev Tankelevitch
Microsoft Research, Cambridge, United Kingdom
Viktor Kewenig
UCL, London, United Kingdom
Auste Simkute
University of Edinburgh, Edinburgh, United Kingdom
Ava Elizabeth. Scott
UCL, London, London, United Kingdom
Advait Sarkar
Microsoft Research, Cambridge, United Kingdom
Abigail Sellen
Microsoft Research, Cambridge, United Kingdom
Sean Rintel
Microsoft Research, Cambridge, United Kingdom
論文URL

https://doi.org/10.1145/3613904.3642902

動画
BIDTrainer: An LLMs-driven Education Tool for Enhancing the Understanding and Reasoning in Bio-inspired Design
要旨

Bio-inspired design (BID) fosters innovative solutions in engineering by drawing inspiration from biology. Learning BID is crucial for developing multidisciplinary innovation skills of designers and engineers. While current BID education has attempted to enhance learners' understanding and analogical reasoning skills in BID, it often relies much on teachers' expertise. When learners turn to learn independently through some educational tools, there are challenges in understanding and reasoning practice in such complex multidisciplinary environment, as well as evaluating learning outcomes comprehensively. Addressing these challenges, we introduce a Large Language Models (LLMs)-driven BID education method based on a structured ontology, as well as three strategies: enhancing understanding through LLMs-enpowered "learning by asking", assisting reasoning by providing hints and feedback, and assessing learning outcomes through benchmarking against existing BID knowledge. Implementing the method, we developed BIDTrainer, an interactive BID education tool. User studies indicate that learners using BIDTrainer understood BID cases better, reason faster with higher interactivity than the baseline, and BIDTrainer assessed the learning outcomes consistent with experts.

著者
Liuqing Chen
Zhejiang University, Hangzhou, China
Zhaojun Jiang
Zhejiang University, Hangzhou, Zhejiang, China
Duowei Xia
Zhejiang University, HangZhou, Zhejiang, China
Zebin Cai
Zhejiang University, Hangzhou, Zhejiang, China
Lingyun Sun
Zhejiang University, Hangzhou, China
Peter Childs
Imperial College London, London, United Kingdom
Haoyu Zuo
Imperial College London, London, United Kingdom
論文URL

https://doi.org/10.1145/3613904.3642887

動画
Teachers, Parents, and Students' perspectives on Integrating Generative AI into Elementary Literacy Education
要旨

The viral launch of new generative AI (GAI) systems, such as ChatGPT and Text-to-Image (TTL) generators, sparked questions about how they can be effectively incorporated into writing education. However, it is still unclear how teachers, parents, and students perceive and suspect GAI systems in elementary school settings. We conducted a workshop with twelve families (parent-child dyads) with children ages 8-12 and interviewed sixteen teachers in order to understand each stakeholder's perspectives and opinions on GAI systems for learning and teaching writing. We found that the GAI systems could be beneficial in generating adaptable teaching materials for teachers, enhancing ideation, and providing students with personalized, timely feedback. However, there are concerns over authorship, students’ agency in learning, and uncertainty concerning bias and misinformation. In this article, we discuss design strategies to mitigate these constraints by implementing an adults-oversight system, balancing AI-role allocation, and facilitating customization to enhance students’ agency over writing projects.

著者
Ariel Han
UC Irvine, Irvine, California, United States
Xiaofei Zhou
University of Rochester, Rochester, New York, United States
Zhenyao Cai
University of California Irvine, Irvine, California, United States
Shenshen Han
University of California, Irvine, Irvine, California, United States
Richard Ko
UCI, Irvine, California, United States
Seth Corrigan
UC Irvine, Irvine, California, United States
Kylie A. Peppler
University of California, Irvine, Irvine, California, United States
論文URL

https://doi.org/10.1145/3613904.3642438

動画
Teaching Middle Schoolers about the Privacy Threats of Tracking and Pervasive Personalization: A Classroom Intervention Using Design-Based Research
要旨

With the pervasive and evolving use of tracking and AI to make inferences about online platform users, it has become imperative for adolescents---a key demographic using such platforms---to develop a deep understanding of these practices to protect their privacy. Traditionally, K-12 cybersecurity education has largely been confined to extracurricular activities, limiting underrepresented students' access. To resolve this shortcoming, we partnered with a rural-identifying middle school to deliver AI-related privacy education in classrooms. Using Design-Based Research methodology, we identified students' AI-related privacy learning needs and developed six education modules. This paper focuses on the design, classroom implementation, and evaluation of module \#2, covering the privacy threats of Tracking and Pervasive Personalization (TaPP). Student assessment outcomes show they developed transferable foundational knowledge of the privacy implications of tracking and personalization after participating in the TaPP module. Our findings demonstrate the benefits of integrating AI-related privacy education into existing K-12 curricula.

著者
Sushmita Khan
Clemson University, Clemson, South Carolina, United States
Mehtab Iqbal
Clemson University, Clemson, South Carolina, United States
Oluwafemi Osho
Clemson University, Clemson, South Carolina, United States
Khushbu Singh
Clemson University, Clemson, South Carolina, United States
Kyra Derrick
Clemson University, Clemson , South Carolina, United States
Philip Nelson
Pandemic Response Accountability Committee, Westminster, South Carolina, United States
Lingyuan Li
Clemson University, Clemson, South Carolina, United States
Emily Sidnam-Mauch
Clemson University, Clemson, South Carolina, United States
Nicole Bannister
Clemson University, Clemson, South Carolina, United States
Kelly Caine
Clemson University, Clemson, South Carolina, United States
Bart Knijnenburg
Clemson University, Clemson, South Carolina, United States
論文URL

https://doi.org/10.1145/3613904.3642460

動画
Putting Things into Context: Generative AI-Enabled Context Personalization for Vocabulary Learning Improves Learning Motivation
要旨

Fostering students' interests in learning is considered to have many positive downstream effects. Large language models have opened up new horizons for generating content tuned to one's interests, yet it is unclear in what ways and to what extent this customization could have positive effects on learning. To explore this novel dimension, we conducted a between-subjects online study (n=272) featuring different variations of a generative AI vocabulary learning app that enables users to personalize their learning examples. Participants were randomly assigned to control (sentence sourced from pre-existing text) or experimental conditions (generated sentence or short story based on users’ text input). While we did not observe a difference in learning performance between the conditions, the analysis revealed that generative AI-driven context personalization positively affected learning motivation. We discuss how these results relate to previous findings and underscore their significance for the emerging field of using generative AI for personalized learning.

著者
Joanne Leong
MIT, Cambridge, Massachusetts, United States
Pat Pataranutaporn
MIT, Boston, Massachusetts, United States
Valdemar Danry
MIT, CAMBRIDGE, Massachusetts, United States
Florian Perteneder
Independent, Hagenberg, Austria
Yaoli Mao
Columbia University, New York, New York, United States
Pattie Maes
MIT Media Lab, Cambridge, Massachusetts, United States
論文URL

https://doi.org/10.1145/3613904.3642393

動画