Teacher Dashboards supported by AI-generated Feedback

JTELSS logo large
Teacher Dashboards supported by AI-generated Feedback Thursday 08/06 10:30-12:00h Plenary Hall Abstract As AI becomes more mainstream, we see opportunities to improve education with its assistance. At the same time, we face the challenge of implementing these innovations in a meaningful way that takes learning science and ethical aspects into

Speakers

Onur Karademir
DIPF | Leibniz Institute for Research and Information in Education, Germany
Daniele Di Mitri
DIPF, Germany
Sebastian Gombert
DIPF, Germany

Start

08/06/2023 - 10:30

End

08/06/2023 - 12:00

Teacher Dashboards supported by AI-generated Feedback

Thursday 08/06 10:30-12:00h
Plenary Hall
Abstract

As AI becomes more mainstream, we see opportunities to improve education with its assistance. At the same time, we face the challenge of implementing these innovations in a meaningful way that takes learning science and ethical aspects into account when implementing these systems, which aim to empower teachers to provide the best possible teaching. In this workshop, the Learning Analytics Cockpit (LA Cockpit), a tool developed by us, will be presented and optimized by the participants. It is a teacher dashboard that allows them to view the learning status of students through different visualizations and to interact with these visualizations to send feedback to (individual or groups of) students to “steer” the learning process. Participants will have the opportunity to test the feedback system and brainstorm requirements for an AI-powered feedback system to enhance the LA Cockpit under the aspects of Feedback Theory, Natural Language Processing/Large Language Models, Recommender Systems,…). After discussing our results, we reflect on how AI systems and Teachers can corporate to push education forward.


Needs Analysis

As AI becomes more mainstream with the help of Chatgpt, we see great opportunities to improve education with AI assistance systems. At the same time, we also face the challenge of implementing these innovations in EdTech tools in a meaningful way, taking into account didactic and ethical aspects, in order to empower teachers to provide the best possible teaching. In this workshop, PhD students will see in a very hands-on way how teachers in school practice use a learning analytics feedback tool, and how it can be improved through AI. Furthermore, they can reflect on how AI technologies and the human touch can co-exist in education.


Learning Objectives

The main Learning Goal for this 90-minute sessions is for the participants to think of these questions:

  • How can Learning Science be incorporated into AI (assistance) Systems?
  • What is the right balance between AI efficiency and human control in education?
  • How we can combine the best of Humans & AI?

Furthermore, the participants should get insights into..

  • How to use Learning Analytics designed to provide feedback to students
  • How such a tool is used by teachers in school practice.
  • Experiencing a Co-Design Session with interdisciplinary Teams (Team AI & Team Learning Science) and define requirements for an EdTech tool


Pre-activities

You should be interested in TEL/EdTech or Learning Analytics. It would also be helpful to be familiar with “The power of feedback” by Hattie, J., & Timperley, H. (2007). This workshop is open to both doctoral students with an educational focus and doctoral students with a technical focus. We need both perspectives!


Session Description

In this workshop, the Learning Analytics Cockpit (LA Cockpit), a tool developed by us, will be presented and optimized by the participants. The LA Cockpit is a teacher dashboard that allows teachers to view the learning status of students through different visualizations and to interact with these visualizations to send feedback to (individual or groups of) students through a feedback system to “steer” the learning process. After a short intro, the tool with its visual feedback system will be tested “hands-on” by all participants in the workshop. Afterwards, we will split into two groups: Group Learning Science & Group AI. Here the groups will think up design suggestions for the existing feedback system of the LA Cockpit under the aspects of Learning Science (possible topics: Feedback Theory, Constructivism) or AI (possible topics: Generative Language Models, Recommender Systems). Afterwards, we will discuss our results and how we can combine the best of the two worlds.