Charting the Future of Education: AI Integration in Aotearoa by 2029
Strategies, challenges, and pathways to transform learning through ethical AI implementation
The New Zealand education system is falling behind other countries in the Asia-Pacific region when it comes to adopting and integrating AI to support learning outcomes. The slow progress can be attributed to various reasons, but the main barrier is the need for more leadership, guidelines, and policy at a Ministry level. This is evident through the lack of documentation compared to countries like Australia, Singapore, and many other countries in Europe. Conversations about AI in education are also not happening to the depth or level required to enable the safe and ethical use of generative AI in schools. Therefore, it was with a mixture of excitement and relief to be invited to the AI forum working group to discuss a blueprint that could direct New Zealand over the next five years with some tangible, immediate actions. Reflecting on thoughts leading up to the hui, discussions during the event and the literature I have been reading lately, this post aims to cover where we could be heading. These views don't necessarily reflect the group and an official paper will be published; however, I thought I'd take the opportunity to write down my thoughts on the flight home.
A Vision of Education in 5 Years with Successful AI Integration for New Zealand
I recently read a position paper by Ouyang and Jiao (2021) discussing three AI paradigms in education. Interestingly, this was published before Chat-GPT was unleashed on the public. However, the three suggested paradigms seem helpful in examining potential visions of the future of AI in education (Table 1).
The role of AI and interactions with the learner distinguish the three paradigms. In Paradigm One, AI directs the learning process, and the learner acts as a recipient. In Paradigm Two, AI supports learning, and the learner collaborates with the system. Paradigm Three views learners as leaders of their own learning, empowered by AI to personalise their learning experience.
While the large tech companies will see easy solutions to push content to the learner in automated systems (paradigm one), we need to look beyond this to maximise AI's potential in education (paradigms two and three). Below are potential scenarios that we could see in five years.
Paradigm 1: Personalised Learning Journeys: AI could enable the creation of customised learning pathways for every student. Imagine AI-driven digital textbooks, like those being developed in South Korea, adapting in real-time to a student's learning pace and style. In this future, instead of one-size-fits-all lessons, students in the same classroom could work on different tasks and activities based on their individual needs, as determined by AI. Effective differentiation is one of the overwhelming challenges faced by teachers each day, and this could be a valuable solution.
However, we need to be careful here that this does not turn into an amplified factory line of education with a soulless robot pushing learners to more of a 'qualification'. You just have to look at the Google reviews of online learning platforms like this to see how much students dislike this teacherless approach - how can this be done differently? Can we avoid AI being used to amplify bad practices in education? This aligns more with what I think AI can do well and quickly (creating content), rather than being critical of what it actually could do to support educators and students in an age of AI.
Paradigm 2: AI - A Collaborative Partner in and out of the Classroom: AI could transition from a technological tool to a collaborative partner in education. The sources from WEF and the OECD suggest that AI could play a crucial role in:
Empowering Teachers: AI could take over administrative burdens, like grading and lesson planning (p.10), allowing educators to focus on providing personalised support and mentorship and fostering meaningful relationships with their students (p.3). This shift could redefine the teacher's role, requiring them to become skilled facilitators of AI-powered learning experiences.
Transforming Assessment: Continuous, AI-driven assessment could replace high-stakes testing. AI systems could be designed to track student progress in real-time, providing instant feedback, and allowing immediate adjustments (p.11) to learning pathways. This could lead to a more holistic understanding of student learning.
Breaking Down Barriers: AI-powered language translation and adaptive content could create truly global classrooms (p.16), connecting students and educators from different countries and cultures. This type of technology could make quality education more accessible to learners in underserved communities and those with disabilities.
This approach requires a greater understanding of how the tools can benefit teaching and learning. Rather than just receiving AI-generated content, the teacher and the student need to build skills in effectively using the tools for learning. Our education system needs to change for this to happen, as it requires us to dismantle assessment and look at what students are really learning at school, and why.
Paradigm 3: AI-augmented learning: This is where it takes a bit more imagination (and trust), and I'm still not entirely comfortable with this potential future. Students and teachers could opt into a full-time personal assistant that they have agency over, understanding how and when to interact with it, bringing it in to augment their learning and their lives.
If we are going to have this level of learner-augmented AI, it is going to need a great deal of data (think of the quantified self - perhaps an extended version of your Garmin 'morning report' or 'Apple Health' with a built-in AI assistant that shares life/learning/work updates and advice based on the conversations that you have combined with digital interactions). However, who will be ready to hand that data to the large faceless companies that own and run LLMs? This is amplified when considering learner data and the almost immeasurable ethical considerations needed to keep young people's data safe. Do we have the capacity to develop something bespoke in New Zealand that we could trust?
Creating the environment for change
For any of these paradigms (particularly two and three) to be realised in the next five years, we should focus on the following areas;
Closing the digital divide: We know a huge gap exists in New Zealand. For any steps to be taken forward, we must commit to narrowing the gap between those with access to technology and the skills to use it and those without.
A Focus on Responsible and Ethical AI Development: This is critical, and the importance of ethical considerations in developing and implementing AI in education should not be underestimated. We have seen that the ethics of the main companies leading the AI charge do not always align with what we expect from companies with so much of our data - see Scarlet Johansen vs OpenAI. In particular, we need to focus on;
Addressing Bias and Ensuring Equity: AI systems should be designed and used responsibly to avoid perpetuating or exacerbating existing societal biases. This requires diverse teams of developers, ongoing monitoring for bias, and clear guidelines for ethical AI use (p.128).
Prioritising Data Privacy and Security: Protecting student data is paramount. Robust data privacy and security protocols would be crucial, ensuring transparency about data collection practices, obtaining informed consent, and safeguarding sensitive information (p.12).
Essential Skills for an AI-Driven World: Successful AI integration would require a greater emphasis on AI literacy for all. This would involve:
Equipping students with future-ready skills: Curricula must prioritise skills like critical thinking, problem-solving, digital literacy, ethical reasoning, and creativity (p.9). AI could be incorporated into all subjects, not just STEM fields, helping students develop a nuanced understanding of its applications and implications.
Preparing educators to thrive in an AI-enabled environment: Robust training and ongoing support would be essential for teachers to effectively integrate AI into their pedagogical practices. This includes developing their confidence in using AI tools, understanding ethical considerations, and adapting their teaching strategies to leverage AI's capabilities (p.12).
A Call for Collaboration and Ongoing Evaluation: All sources, policies and guidelines that I have read stress that successfully integrating AI in education requires a collaborative, human-centred approach, where:
Educators, policymakers, researchers, and technology developers work together (p.4) to create AI-powered tools and resources that are pedagogically sound, ethically grounded, and meet the needs of diverse learners.
Ongoing research and evaluation are crucial to understanding AI's impact on learning outcomes, student and teacher well-being, and the broader educational landscape.
Of course, the successful integration of AI in education is not without its challenges; addressing issues like potential job displacement, ensuring equitable access to technology, and navigating the ethical complexities of AI will be crucial. Understanding the broader impacts of mass energy consumption, the worrying reports of AGI, and the downfall of the human race should also be placed within any discussions regarding the future of AI in education. However, the potential benefits of AI, if implemented thoughtfully and responsibly, could revolutionise education, creating more personalised, engaging, and equitable learning experiences for all.
At the blueprint working group we were also tasked with developing several key tasks that could be implemented in the next year to support change.
Immediate Action…
We can take the next steps over the coming year to enable the thoughtful and effective integration of AI in education.
Prioritise Infrastructure: Education systems should prioritise equitable access to affordable, high-quality connectivity and digital learning resources. This includes ensuring all learners and educators have access to reliable internet connectivity and digital devices at school and at home. This is essential for the equitable use of AI-powered tools and resources.
Develop Clear Guidelines and Policies: There is an immediate need for education systems to establish clear guidelines and policies to ensure the ethical and effective integration of AI. This includes:
Addressing data privacy concerns: Develop and implement robust data privacy and security protocols to protect sensitive student information. This includes ensuring consent for data collection, anonymising data, and limiting data collection to only what is necessary for educational purposes.
Promoting ethical AI use: Immediately establish ethical guidelines for AI use in schools that consider potential bias, transparency, accountability, and the impact on student and teacher well-being. This can be achieved by adapting existing ethical frameworks, such as the soon-to-be-released AI Competency Framework for Teachers by UNESCO, to suit specific national contexts.
Creating a culture of responsible AI use: Encourage teachers, students, and other stakeholders to participate as co-designers in the research and development of AI-enabled educational tools. This collaborative approach will help ensure these tools are pedagogically sound, relevant, and used effectively.
Support Teacher Development: We should focus on building teacher capacity and confidence in using AI for teaching and learning. This can be achieved by:
Integrating AI into teacher training: Initial teacher education programs should incorporate training on the use of digital tools for learning across all subjects.
Providing ongoing professional development opportunities: Continuous professional development for teachers should include training on the use of AI in teaching, learning, and assessment. A community of practice involving interested and passionate educators could drive this (join here!).
Fostering AI literacy: Develop and implement government-endorsed AI curricula that cover ethical considerations, understanding algorithms, and the proper use of AI tools. This will help equip both teachers and students with a foundational understanding of AI.
As we slowly integrate AI into education, it becomes clear that the steps we take must go beyond adopting new technologies or succumbing to 'AI solutions' imposed by big tech companies. We need a responsible and ethical approach prioritising human-centred learning experiences which can celebrate and enhance educators' skills. With this in mind, charting the future of AI integration in New Zealand education requires us to move beyond paradigm one, where AI merely automates and directs learning, risking education becoming solely a transactional pursuit of qualifications and credits.
Processing these thoughts and understanding more about the potential paradigms prompted me to revisit Gert Biesta's domains of purpose —qualification, socialisation, and subjectification, which provide a framework to transcend the possible limitations; we risk being trapped in the qualification domain if we only focus on paradigm one, content pushed to speed up the process of gaining credits. By integrating AI with a focus on socialisation and subjectification alongside qualification, we can cultivate critical thinking, ethical reasoning, and meaningful societal contributions in our students. This holistic approach ensures that AI enhances education by preparing students not only for personal success but also for active and positive participation in shaping our shared future.
This post was written by me, a Human, I used NotebookLM to help synthesise the sources mentioned in this post. I’ve tried a different approach to referencing the links by using the hyperlink and then the page number with the relevant reference - I hope that is helpful.
Very interesting and challenging article. Thanks Tim. A question: Tim what's your background and especially your teaching experience and areas of academic subject matter?