1 of 27

AI Literacy Framework

Associate Professor Kathryn MacCallum - University of Canterbury

Dr David Parsons - academyEX

Dr Mahsa Mohaghegh - AUT

2 of 27

“Tino rangatiratanga: mō tātou, ā, mō kā uri ā muri ake nei”

The ability to create and control our destiny for generations to come

This is the whakataukī that means most to Māori data, AI and technology ethicist Dr Karaitiana Taiuru (Ngāi Tahu, Ngāti Kahungunu, Ngāti Toa). It reminds Taiuru why he still does what he does - and that there is a new generation of children who will not know life without AI and it is "our role to create the new normal for them".

RNZ. (2024, March 18). Māori AI expert Dr Karaitiana Taiuru shares his favourite whakataukī. https://www.rnz.co.nz/news/national/512062/maori-ai-expert-dr-karaitiana-taiuru-shares-his-favourite-whakatauki

3 of 27

Session Agenda

What is AI Literacy?
What’s in our AI Literacy Framework
How can it be applied?
How can you contribute?

4 of 27

Addressing AI Literacy

The current wave of AI technologies has added a sense of urgency to emerging questions about AI literacy - the set of skills and competencies that are needed for everyone (but perhaps especially educators and learners) to successfully adapt to this new environment

We will explore what AI literacy is, why it matters, and how we might develop it.

5 of 27

What is AI Literacy?

Initially conceptualised by Yoko Konishi in 2016 as being able to recognise tasks that can be performed by AI, and learning and investing in the human strengths that it cannot replace

Since then, many others have suggested various other definitions. One commonly referred to definition is Long and Magerko’s (2020) five AI literacy themes:

What is AI?
What can AI do?
How does AI work?
What should AI do?
How do we perceive AI?

6 of 27

How’s Your AI Literacy?

Please respond using this Google Form to a few questions designed to assess a few aspects of AI literacy

academyEX

7 of 27

What is an AI Literacy Framework?

Having some ideas about what AI Literacy might involve is not enough to structure and evaluate how we can develop AI literacy
An AI literacy framework can take many forms but its main objective is to provide an organised way of thinking about how someone might scaffold their learning about AI

“An AI literacy framework is a structured approach to understanding, teaching, and learning about artificial intelligence (AI). It is designed to provide a comprehensive set of guidelines, competencies, and knowledge areas necessary for individuals to effectively engage with and understand AI technologies.” - ChatGPT

Some measure of progression is important, though often missing from existing frameworks

8 of 27

The AI Literacy Research Project

The AI literacy framework was developed through a ‘Delphi Study’ refined over three rounds

Panel of 17 AI experts (education = 11, industry = 5 and 1 with both)
Expertise in a range of topic (AI ethics, system building, probabilistic models, general and specialist AI teaching, Māori cultural AI ethics, privacy, cybersecurity, online safety)
Educational Experts (tertiary educators = 4, school teachers/principals: 3, support teaching roles: 4 (i.e lab technicians, learning support, designers)
Industry Experts had a variety of roles (incl. Senior Quality Editors, Senior Principal Data Scientists, Special AI Projects Leads, Business Founder)

You can find more details on this web page

academyex.ac.nz/ai-literacy-research-project/

9 of 27

AI Literacy Framework Levels

10 of 27

AI Literacy Framework Categories

The six categories of the framework are divided into three domains of literacy.
These apply at all four levels, but in each case are further developed to address the appropriate level of capability.
These categories together ensure that there is an appropriate mix of knowledge, skills, and critical thinking.

academyEX

11 of 27

Learning Areas with the Framework - Level 1

In applying the framework to specific learning tasks, a range of different topics and activities can be extracted to address the various components at each level

12 of 27

Learning Areas with the Framework , Level 2

13 of 27

Learning Areas with the Framework , Level 3

14 of 27

Learning Areas with the Framework , Level 4

15 of 27

AI or Not? Quiz

What is AI and what isn’t is actually very complex especially as there is huge debate around what can be termed AI (most of what is AI is narrow AI). This Kahoot! Is a fun way to explore students’ understanding

Shared version of this Kahoot!

All Ages

This article talks more about why AI is not actually intelligent.

https://dayofai.org/ also has some fantastic resources to explore

academyEX

16 of 27

Quick, Draw!

Quick, Draw! provides real-time feedback as you draw. Students can see how good the neural network is at recognising their drawings. It also makes its data set publicly available, meaning that other researchers can use the data to train their own neural networks.

All Ages

https://quickdraw.withgoogle.com/

academyEX

17 of 27

Code.org AI for Oceans

An interactive game teaching young children about machine learning (ML). It teaches children about how AI is trained through pattern recognition. It explores the importance of correct training and the biases that can be introduced in training.

https://studio.code.org/s/oceans

Primary

academyEX

This example can be used to explain level 1 and 2 levels:

The Impacts of AI:

Level 1: Describe how AI is impacting society

Level 2:

Discuss the role of humans in the development, deployment, and ethical use of AI
Demonstrate an understanding of AI's role in a specific context
Explain the factors in AI adoption related to AI usage and potential harms

What AI Is & How It Works:

Define the different types of AI
Explain the ways data is used in different AI systems
Explain how users can interact with AI systems (including using NLP)
Be able to identify different ways that AI is being used within different applications/contexts
Define the fundamental terminology of AI

Cognitive Skills:

Level 1:

Evaluate the role of data within AI systems and the implications data has on training of AI models
Identify the affordances of AI technologies for different contexts
Assess the implications of using AI for a specific purpose/context

Level 2:

Discuss how AI can empower people to make better decisions and augment their intelligence
Discuss issues related to the understandability of AI processes, such as AI transparency and explainable AI

Social & Ethical Issues:

Level 1:

Assess the potential benefits to society of the use of AI
Identify ethical implications of AI (e.g., bias, fairness, transparency, accessibility, and accountability)

Level 2:

Understand how bias occurs in AI systems
Discuss the social impacts of AI, including its effects on societal norms and cultural biases

Challenges & Opportunities

Level 1:

Discuss how AI can impact cultural diversity, recognising biases in AI systems

Level 2:

Examine risks related to use of data in AI systems, including data collection, storage, and usage
Discuss how AI systems can avoid reinforcing disadvantage for certain groups

18 of 27

Teachable Machine

A hands-on tool to learn about supervised ML. Students can create their own AI systems to detect images, sounds and even poses. They learn the importance of having the right training data and the difference between training and testing data.

https://teachablemachine.withgoogle.com/

BONUS: For a ready to use lesson plan by MIT

All Ages

academyEX

19 of 27

How LLMs are trained (unplugged)

This unplugged demonstration gives a simple example of how LLMs work, highlighting the role of probability using a Dr Seuss story

All Ages

Slideset

Video example

EdPuzzle Version

academyEX

20 of 27

AI Bingo (Unplugged)

Students are given bingo cards with various AI system examples. Students find a partner who has also used that AI system and work together to identify what prediction the system is making and the dataset it uses.

All Ages

Slides for AI Bingo

Teacher resource

academyEX

21 of 27

Design Evaluation Using the Framework

The provided domains and categories can be used to evaluate course materials at the appropriate level, or to design new learning experiences.

For level 1, there is an online tool that can help to analyse a course design and give structured feedback on possible improvements to AI literacy coverage

Examples of foundational AI literacy for everyone:

1. Basic understanding of AI concepts:

- What AI is and isn't

- Common AI applications in everyday life (e.g., voice assistants, recommendation systems)

- Simple explanation of machine learning

2. Awareness of AI's impact:

- How AI affects decision-making in various fields (e.g., healthcare, finance)

- Potential benefits and risks of AI in society

3. Critical thinking about AI:

- Recognizing AI-generated content

- Understanding the limitations of AI systems

4. Ethical considerations:

- Basic awareness of AI bias and fairness issues

- Privacy concerns related to AI and data collection

Developing AI literacy can be tailored to different learners and contexts

1. **Primary Education**: Introduce AI literacy using simple examples and interactive activities. For instance, explain how virtual assistants work or how recommendation algorithms function. Use tools like Scratch for hands-on experience.

2. **Secondary Education**: Dive deeper into AI concepts like machine learning and data ethics. Students can build simple AI models using tools like TensorFlow Playground. ![TensorFlow Playground][Teachable Machines ]

3. **Higher Education**: Engage with complex AI concepts and explore ethical, social, and economic implications.

4. **Workplace Training**: Professionals can learn about AI applications specific to their industry through workshops or online courses.

22 of 27

Your Feedback

We would welcome your feedback on the AI Literacy Framework

You can share your responses in this Google Form

There are only three questions but it would take some time to answer them all. We welcome your feedback, however incomplete.

academyEX

23 of 27

QnA

Dr Kathryn MacCallum

University of Canterbury kathryn.maccallum@canterbury.ac.nz

Dr David Parsons

academyEX david.parsons@academyex.com

Dr Mahsa Mohaghegh

AUT

mahsa.mohaghegh@aut.ac.nz

Copy of the slides

academyEX

24 of 27

Level 1

The first level of the framework provides the foundational level of AI literacy for everyone.
How and when these literacies might be developed will vary across learners and contexts and may be developed in different sequences and at different depths.

academyEX

25 of 27

Level 2

The second level of the framework provides a more active approach to AI literacy.
Building on the foundational skills of level 1, it provides the learner with a pathway to become directly involved in the use and application of AI tools and engage more critically in the wider issues around AI technologies.

academyEX

26 of 27

Level 3

This level of the framework provides the highest level of knowledge skill, and competency within the scope of AI literacy
Comprehensive understanding of AI concepts
Utilising AI tools to build applications
Working with AI in a principles based, ethical and inclusive way

academyEX

27 of 27

Level 4+

Level 4+ suggests what comes beyond literacy
Advanced concepts
Research
Strategy and Policy development
AI Systems development

academyEX