Introduction to Artificial Intelligence 

The field has a long history and appears in many areas of our lives. AI technology can be found in familiar digital tools, like autocomplete texting or recommended content on social media, as well as emerging tools, like driverless cars and image generators. The possibilities of AI tools are broad, but they can be understood through some common terms used in the field. This factsheet outlines some of these terms. For more details and examples, see the Exploring further section at the end of this resource.  

Individual pieces of AI software are often referred to as models or systems. The difference between the AI models we use today and those seen only in science fiction can be understood through the ideas of weak and strong AI. 

The field of AI contains many related subfields and concepts. Some of these are illustrated here and explained further below. 

• Weak AI describes all AI models in use today. It refers to systems designed for specific tasks, like virtual assistants, chatbots, recommendation systems, and autonomous vehicles. Weak AI excels at specific functions but lacks general intelligence.  
• Strong AI represents theoretical AI models with human-level intelligence or capabilities surpassing human intellect, and the ability to apply this to a wide range of tasks. Strong AI does not currently exist, and there is debate around whether it may come to exist in our lifetime.  
• Machine learning is a subfield of AI that involves creating systems that adapt with minimal human input. By analysing large amounts of data, machine learning software can identify patterns and relationships to make predictions or generate responses. Deep learning is a specialised form of machine learning that relies on artificial neural networks inspired by the human brain. It enables systems to learn and improve from experience with less explicit programming.  
• Generative AI models create new content based on training from existing data. These models use complex algorithms to generate text, images, or even video and music based on patterns identified through machine learning. Generative AI has gained attention from the public, businesses, and governments for its ability to create human-like outputs. You may be familiar with generative AI models such as ChatGPT, Bard, and DALL-E.  
• Natural Language Processing (NLP) involves the processing and understanding of human language by computers. It involves the development of models that enable computers to process, interpret, or generate human language in a way that can be meaningful and useful – encompassing tasks such as language translation, text summarisation, and question-answering systems.  
• Large Language Models (LLMs) are generative NLP systems that use deep learning to process huge amounts of data. LLMs like ChatGPT and Bard are trained on vast amounts of written work and generate human-like responses to conversational-style prompts. This creates new ways to interact with machines: even if a piece of software doesn’t produce text, an LLM could let it process prompts in conversational human language.  

Data sovereignty  

Many AI models are trained on huge amounts of data produced by people, and there are many ongoing discussions about the power people should have in deciding how their data is used. Decisions around ownership and the fair use of data are very important across the field of AI. In Aotearoa New Zealand, Māori data sovereignty principles are a significant consideration under Te Tiriti o Waitangi. These principles emphasise that Māori data should be under the control of Māori. Collaboration between Māori communities, the government, and technology stakeholders is essential to establish frameworks that protect Māori data sovereignty, preserve cultural integrity, and mitigate potential risks. 

There are many other branches of AI which are not covered in this resource. These branches form a diverse and interconnected field that collectively works towards building ‘intelligent’ systems and solving complex problems. Each branch brings its own set of techniques, methodologies, and challenges, contributing to the broader field of AI. 

Reflection prompts 

• When people hear the term artificial intelligence, do you think they mostly imagine something closer to strong or weak AI? How might this image affect their feelings towards AI?  
• How comfortable are you with discussing new technologies in your teaching? Do you discuss why people might choose to use – or not use – new technologies? 

AI technology offers opportunities across many sectors of society. As you explore and discuss these opportunities, note that many of them come with challenges as well. Compare this information with the challenges factsheet to draw connections between the two. 

• Data-driven insights AI-powered analytics can analyse vast amounts of data to generate valuable insights. For example, AI has been applied to analyse aerial footage to support disaster response in the South Pacific. (Using AI For Good: A New Data Challenge To Use AI To Triage Natural Disaster Aerial Imagery)
• Efficient administrative tasks AI can automate administrative tasks, freeing up time for other duties.  
• Intelligent tutoring systems AI can support individuals in their learning journey by providing real-time feedback, guidance, and personalised recommendations.  
• Augmented creativity and innovation AI tools can foster creativity by rapidly exploring new ideas or providing alternative perspectives.  
• Access to resources AI can make access to information more inclusive and equitable. For example, some generative AI models can produce captions and audio descriptions for images and video, enabling people to access this content using a range of senses. (How Artificial Intelligence is Improving Digital Accessibility)
• Language translation and accessibility AI-powered language translation and accessibility tools can help bridge language barriers and improve access to information for diverse populations. For example, natural language processing AI could make it easier for digital tools to understand and present speech and text in te reo Māori, empowering its place in the digital world. (Accelerating the revitalisation of te reo Māori with AI)
• Fraud detection AI algorithms can analyse large volumes of data, such as transaction records or user behaviour patterns, to detect fraudulent activities.  
• Predictive maintenance AI can be used to analyse data from sensors and equipment to predict when maintenance is needed, allowing for proactive maintenance scheduling and reducing downtime. 

AI and Industry 

In commerce and industry, AI can revolutionise processes, increasing efficiency and productivity. Through AI-powered automated systems that can analyse vast amounts of data, organisations can make informed decisions and gain valuable insights. This means AI-powered technologies have the potential to transform industries such as healthcare, transportation, and manufacturing, where they can assist in diagnosing diseases, optimising logistics, and streamlining production processes. These advancements hold the promise of improved services, economic growth, and higher standards of living. 

Careers in AI 

While computer science-based careers are prominent within AI, it’s important to note that AI expertise is increasingly valuable across diverse fields. Problem solvers who possess an understanding of AI can contribute to developing innovative solutions, while policymakers and lawmakers need knowledge of AI to effectively regulate its ethical and legal implications. It is anticipated that in the future, most professions will incorporate some form of AI technology.  

The broadening scope of AI opens up a wide range of potential career paths for students. By acquiring the necessary skills and knowledge in AI, individuals can position themselves to thrive in a rapidly growing industry. 

Reflection Prompts 

• Do your programmes of learning involve thinking about the future, and how big problems could be solved? How might AI factor into this thinking?  
• Do you discuss particular careers or industries in your teaching? How might AI affect these?  
• New Zealand society has both influenced, and been influenced by, technological innovation. What opportunities presented by AI technologies seem the most valuable to embrace? 

The opportunities of AI technologies come with a range of challenges that must be actively navigated by those who develop and use these technologies. Consider how some of these challenges may not apply in some applications of AI, but may be of high concern in others. 

• Privacy concerns AI tools may collect and store sensitive personal data, posing security and privacy risk.  
• Plagiarism risks If generative AI is used to produce writing or images without being credited, this could be considered plagiarism.  
• Ethical sourcing and cultural appropriation The responsible acquisition of data and avoiding cultural misappropriation can be complex when using AI tools. Some AI models may be trained on data created by people who have not been appropriately consulted, credited or compensated. This is a critical data sovereignty consideration. For example, some AI-powered apps have animated images of Māori tīpuna for entertainment, breaching tapu. (Te Ao Māori considerations of AI with the dead and personal Data)
• Terms of use Many AI tools have terms of use that include age limits, which should be considered when using with young people.  
• Emotional support limitations AI lacks the ability to provide the same level of emotional support and understanding as human interactions.  
• Limited creativity Many generative AI tools cannot create outputs beyond the limits of the data they are trained on. This constrains their originality.  
• Incorrect information Many generative AI models are designed with the primary goal of providing human-like outputs, rather than accurate information. This means that they can sometimes provide incorrect information in a convincing manner. For example, generative AI poses the risk of erasing Māori leaders and achievements through the false retelling of stories and histories. This can undermine indigenous sovereignty and cultural preservation and revitalisation. (Te Ao Māori considerations of AI with the dead and personal Data) 
• Changing labour demands Alongside new jobs and opportunities, technological developments can change the nature of work and make some roles redundant.  
• Expensive start-up costs Developing and implementing AI-powered tools can require significant financial investment. Even AI tools that are free or cheap for users are often powered by expensive servers.  
• Sustainability The powerful servers used for some AI technology can use large amounts of electricity, which can be environmentally harmful if not sustainably sourced. 

Ethics, bias & morality  

It is important to remember that AI models are made by humans. Humans – whether consciously or unconsciously – hold biases which can be reflected in the technologies they create.  

These biases might take shape in the data that an AI model is trained on, or the algorithm used to process that data (known as algorithmic bias). Biases can emerge because of cultural, social, or institutional expectations. The result is decision-making by computers that are systematically less favourable to certain groups of people. An example of this could be a facial recognition AI model which is not trained on diverse data. This could make the resulting model more likely to misidentify people whose images are not represented in that training data.  

Sometimes human morality is intentionally built into AI systems through training based on moral rules. For example, AI technology is used on some social media platforms to help apply content moderation policies. The AI model could be trained on social media content to help it recognise the patterns associated with bullying behaviour. It could then be used to flag or block content that follows these patterns. However, the rules an AI model follows are reliant on the morals of the humans creating it: different social media companies may train AI to block or allow different types of content.  

By growing an understanding of these ethical challenges, students can become more informed and critical users – and creators – of AI tools. 

Reflection prompts 

• Do you teach about ethics and bias? Could conversations about AI support this teaching?  
• How do you think a person or company should evaluate the risks of using a new AI tool?  
• Could some of these challenges become drivers for positive change? What steps could people take to work towards this? 

Introduction  

Kit Willett, a secondary English teacher and DigiTech team leader at Selwyn College, Auckland, describes himself as “not a tech expert but a tech enthusiast and user.” This case study examines Kit’s journey in learning and teaching about AI. It also outlines some cases in which Kit has used a generative AI tool in his practice as a teacher. For guidance on using generative AI in schools, see the Ministry of Education website. 

 
Inspiration  

Kit has been interested in AI, and especially generative AI, for a while before it “kind of got big”. He has been exploring uses of generative AI through image generation and large language models (LLMs) for the past four years.  

Recently, Kit has been supporting secondary teachers to explore ChatGPT and other AI tools by running professional learning events, and through the publication of his booklet, ChatGPT @ School: A guide to language models for secondary schools.

Implementation 

“Teachers must be enablers of AI skill development.”

Kit believes that to be able to use generative AI like ChatGPT effectively, all kaiako and ākonga need to understand what it is, how it is trained, and how it works. This way they can make discerning decisions about the use of these digital tools in teaching and learning.  

He suggests all learners need to make ethical choices, understand the bias in machine learning and algorithms behind the tool, and gain awareness of the importance of the authentic demonstration of fundamental skills after contextual learning has taken place.  

Kit explains that ChatGPT should not be treated as a primary source of information, as this is not its purpose – its purpose is to manipulate text. “We understand the world through language, but it does not understand the world. It can present information and support research, but it is not a proxy for knowledge.” AI tools do not have human discernment, hence ākonga need an awareness of how it works, to understand when it is appropriate to use it as a tool, and how to evaluate the usefulness of its outputs. Being able to make informed judgements about the content generative AI tools produce is a useful skill. Understanding how these tools work will support students to use them well.  

Te Tiriti o Waitangi must be considered by any teachers who wish to explore using generative AI as a tool for learning. Māori data is a taonga, and teachers need to be aware of Māori data sovereignty to make suitable choices for all our tamariki – AI cannot replace engagement with local whānau, hapū and iwi. 

Outcomes 

After learning and teaching about AI, Kit has applied this knowledge to explore how AI tools can be used in the classroom. He suggests that generative AI is an exciting development to expedite tasks teachers are already doing. For example, an experienced teacher knows how to create successful lesson plans linked to specific curriculum content. By using generative AI, they could speed up the process of content creation if they make sure to review and adapt the AI-generated work. Other time-saving opportunities Kit is exploring include generating learning tasks for students who need extra non-assessed tasks, or supporting the creation of differentiated lessons.  

Kit comments that kaiako need to remember that AI tools like ChatGPT do not know everything – they are conversational text generators. The teacher is the knowledge expert in their classroom and, as such, needs to develop student/teacher relationships which enable learners to feel confident about asking for support.  

However, Kit believes there are benefits for students exploring the use of generative AI – provided they meet the tool’s terms of use. Teachers can encourage learners to be drivers of their own learning by using generative AI to do things like clarifying tasks, brainstorming into contexts, building skills for planning, evaluating the reliability of sources, and providing scaffolds for writing. However, the learning intentions of the task must be made clear to ensure the use of AI doesn’t prevent ākonga from engaging in key learning experiences.  

When using AI in these ways, learners need to understand the challenges associated with them, such as the importance of disclosing ownership of the generated outcomes. For many assessments, students may not utilise generative AI because their work must be created independently, and they need to be able to read, understand, and claim the outcomes as their own work. 

What next?  

“AI has the potential to revolutionise education by empowering teachers with powerful tools and insights that enhance the learning experience, making them even more effective in nurturing the minds of students.” 

The next steps for Kit are to develop, test and trial more learning opportunities for the students in his English classes. One example could be the use of ChatGPT to produce an essay based on an NCEA standard or school-set marking criteria. Then students could read through ChatGPT’s essay, using the marking rubric to objectively assess and develop annotation and evaluation skills.  

Kit has also produced resources to support teachers with their understanding and use of ChatGPT as a tool. A link to these can be found in the Exploring further section of this resource. 

Suggested learning experiences  

• Engage students in a discussion about AI’s impact on society. Choose an area of society relevant to your current programme of learning. Ask students to brainstorm potential applications of AI in this area and explore the associated benefits and challenges.  
• Divide students into small groups and assign each group a specific, publicly accessible AI tool or application. Ask them to research and present how the tool works, its benefits, and any ethical considerations associated with its use.  
• Ask your students to think about AI and their own values. Suggest a possible situation in which AI technology could be used, then ask students to stand on a ‘values continuum’ from “AI should be used” to “AI shouldn’t be used” in the situation. Ask volunteers to explain the thinking behind their position.  
• Ask groups of students to write a question to test for ‘strong AI’. Ask them to think about what makes human intelligence different from what machines can currently do, and use this thinking to justify their question. If possible, ask some of these questions of a conversational AI model to see if it ‘passes’ as intelligent.  

 
Further reading 

• Kit Willett has produced a number of AI-related resources for teachers, which can be found on his website: ChatGPT Resources – Kit Willett
• For a simple explanation of AI concepts in video format, see this five-minute video from the Boston Museum of Science: What is AI? – YouTube 
• See the IBM page on AI Ethics to learn more about ongoing discussions in this area: AI Ethics - IBM   
• To learn more about Māori data sovereignty, see the Te Mana Raraunga website: Te Mana Raraunga 
• For more about how AI models can be trained, see the Machine Learning page by MIT Sloan: Machine learning, explained – MIT Sloan School  
• The field of AI has a long history. Harvard University provides a summary of this history: The History of Artificial Intelligence – Science in the News (harvard.edu)
• The Canterbury University computer science field guide provides interactive resources for high school students learning about computer science: Computer Science Field Guide

1. Using AI to support Pacific disaster recovery: Using AI For Good: A New Data Challenge To Use AI To Triage Natural Disaster Aerial Imagery (forbes.com)
2. AI for accessibility: How Artificial Intelligence is Improving Digital Accessibility
3. Applying AI to support the revitalisation of te reo Māori: Accelerating the revitalisation of te reo Māori with AI – AI for Good 
4. Concerns around use of Māori data: Te Ao Māori considerations of AI with the dead and personal data – Dr Karaitiana Taiuru PhD