AI Automation for Corrective Feedback in Language Learning
Using AI to Automate Corrective
Feedback in the Learner’s L1
Rob Hirschel, Sojo University (hirschel@m.sojo-u.ac.jp)
Chris Tempest, Sojo University (ctempest@m.sojo-u.ac.jp)
18 May, 2024
Reasons for not giving corrective feedback
1.
Time-consuming
2.
Delay in getting feedback to students
3.
Comprehensibility of the feedback in the L2
4.
Students potentially disregarding the feedback
5.
Potential demotivation
Some Existing Programs
New AI Text item created by Marcus Green
•
Available for free on Github
•
Requires AI Connect Tool or AI Connector (both also free)
•
Requires either paid ChatGPT API account (Open AI) which will give access to
ChatGPT4 or Ollama hosted Large Language Model
AI Connect Tool
AI Connector
ollama
Open AI
a trip
For those with
Admin access
Interface in Moodle Quiz
AI Prompts
Student Feedback (Likert-scale Data)
1
(
S
t
r
o
n
g
l
y
d
i
s
a
g
r
e
e
)
~
2
~
3
~
4
~
5
(
S
t
r
o
n
g
l
y
a
g
r
e
e
)
Student Feedback – Comments
(translated unless indicated by
※)
Could clearly understand where I made mistakes
I felt it was easier to use the keyboard (as opposed to writing on paper)
※ Nice :)
I thought that my previous English writing was accurate, so this system is
helpful
Until now I wondered if my sentences were correct, so the ChatGPT
explanations are extremely good
※ Errors check was interesting!!
Where we go from here
•
We are happy with the results so far
•
BUT we need to determine what is the best feedback
•
AND we need to learn how we can refine the prompts
•
This is where we turn to you and ask for your advice
Thank you very much! Please feel free to ask us any questions!
Rob Hirschel (hirschel@m.sojo-u.ac.jp) Chris Tempest (ctempest@m.sojo-u.ac.jp)
Exploring the utilization of AI technology to automate and enhance corrective feedback in language learning, addressing challenges such as time constraints, delayed feedback, and student comprehension. Introduction of a new AI text tool by Marcus Green for accessible and effective feedback provision, leveraging platforms like ChatGPT. Student feedback indicates improved understanding and engagement with the AI-driven feedback system.
Download Presentation
Please find below an Image/Link to download the presentation.
The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author. Download presentation by click this link. If you encounter any issues during the download, it is possible that the publisher has removed the file from their server.
E N D
Presentation Transcript
Using AI to Automate Corrective Feedback in the Learner s L1 Rob Hirschel, Sojo University (hirschel@m.sojo-u.ac.jp) Chris Tempest, Sojo University (ctempest@m.sojo-u.ac.jp) 18 May, 2024
Reasons for not giving corrective feedback 1. Time-consuming 2. Delay in getting feedback to students 3. Comprehensibility of the feedback in the L2 4. Students potentially disregarding the feedback 5. Potential demotivation
New AI Text item created by Marcus Green ollama Available for free on Github Requires AI Connect Tool or AI Connector (both also free) Requires either paid ChatGPT API account (Open AI) which will give access to ChatGPT4 or Ollama hosted Large Language Model Open AI AI Connect Tool AI Connector
For those with Admin access
1 (Strongly disagree) ~ 2 ~ 3 ~ 4 ~ 5 (Strongly agree) Student Feedback (Likert-scale Data) Was the feedback ~ 4.76 4.8 4.75 4.68 4.7 4.61 4.65 4.6 4.52 4.52 4.55 4.47 4.5 4.45 4.4 4.35 4.3 Useful? Mech (#1) Comprehensible? Art & Des (just once) Mech (#2)
Student Feedback Comments (translated unless indicated by ) Could clearly understand where I made mistakes I felt it was easier to use the keyboard (as opposed to writing on paper) Nice :) I thought that my previous English writing was accurate, so this system is helpful Until now I wondered if my sentences were correct,so the ChatGPT explanations are extremely good Errors check was interesting!!
Where we go from here We are happy with the results so far BUT we need to determine what is the best feedback AND we need to learn how we can refine the prompts This is where we turn to you and ask for your advice Thank you very much! Please feel free to ask us any questions! Rob Hirschel (hirschel@m.sojo-u.ac.jp) Chris Tempest (ctempest@m.sojo-u.ac.jp)
