This page contains the resources that have so far been created to support educators and learners as part of this project.
Exploring AI Bias with the MIT Cat Dog Dataset
MIT created An Ethics of Artificial Intelligence Curriculum for Middle School Students, some of which works equally well at tertiary level for developing AI literacy.
One of the activities uses a dataset of dogs and cats that can be used in conjunction with Google’s Teachable Machine to explore bias in machine learning systems.
This resource collates the materials along with a slide set that makes it easy to use this resource with students.
- Here is the slide set, which structures the learning activity
- Here is an individual slide that contains the table where students can fill in the results of their testing
- This zip file contains the images for the initial cat dataset
- This zip file contains the images for the initial dog dataset
- This zip file contains the images of cats and dogs for the test data
- This zip file contains the images for the recurating dataset, which contains more cats and dogs
The AI Literacy Design Analyser
The AI Literacy Design Analyser is a web-based tool that provides an interactive feedback mechanism for assessing the coverage of AI literacy in course design. Based on statements from Level 1 of the AI Literacy Framework, it provides a series of statements about each of the six categories of the framework. The user can select the level of presence, based on a scale of 0-4, of each of these statements in a particular course design. The tool uses these responses to help evaluate the AI literacy content of the design and gives feedback as to how it might be improved. The tool can be used multiple times to help refine the design of a course.
The design of the analyser tool is relatively simple. Likert scale questions have been developed for each of the 24 literacy components in level 1 of the framework. Each input view of the tool presents the questions related to one of the 6 framework categories, with a 5-point slider for each response. At the end of the input process, the data is analysed and the results are presented in the following ways:
- A radar chart showing the extent of AI literacy in the course design
- The least-covered categories are identified
- Suggestions are given as to how the design can be improved
- Text and spreadsheet summary data can be downloaded
Here is an example of a summary chart
And here is some example feedback based on the data related to the chart above.
AI and Authentic Assessment
This slide set provides some ideas for educators who want to make their assessments more authentic (and inclusive) in the context of AI