Connecting AI to educational goals
Educational AI needs clear objectives, led by teachers, to work well
Interviewee:
Interview by:
Rose Luckin has spent more than three decades designing and evaluating technology and AI for education. Rose tells Helena Pozniak about the promise and risks of AI in schools.
Helena Pozniak: You’ve said you are optimistic that we can become smarter in this age of AI – what do you mean?
Rose Luckin: We’re more intelligent today than we were 500 or even 100 years ago. I believe that now we need to increase the pace of change and become smarter by increasing the sophistication of our intellectual activity. We need to develop ever more advanced critical thinking skills, metacognition, and self-directed learning capabilities, for example. We also need to learn how to leverage AI effectively to support our intellectual growth – this will be a game changer.
Part of the secret of success when it comes to AI is recognising that we actually have to work harder intellectually. Commercialisation of AI means that the message is very much about AI making life effortless. But in fact it’s absolutely the opposite. Now we have these smart technologies we must achieve more.
“Part of the secret of success when it comes to AI is recognising that we actually have to work harder intellectually.”
HP: You’ve also said we underestimate our own intelligence?
RL: We need to widen our lens on human intelligence. Our meta-intelligence – our ability to reflect on our physical position in the world with respect to things and people – is profound. Consider the level of meta-intelligence required to do simple tasks such as travel abroad. This is sophisticated but we don’t recognise or value it. Yet an AI couldn’t begin to do that.
HP: How does this relate to education?
RL: Look at it through a developmental lens. My three year old grandson couldn’t travel as I can because he doesn’t yet have an understanding of the world. How can we help learners develop this meta-intelligence? What does it mean to be an effective learner in an educational establishment? How can you improve the way you navigate your physical and digital environment, interact with people and understand your position in this world?
Understanding the difference between human and artificial intelligence will help us make the right decisions about how we use and develop AI. We have a uniquely human ability to reflect on our own thought processes, our own emotions, our own experience of the world. Metacognition is our ability to be aware of our own cognition, being able to plan, monitor, regulate and reflect. These meta-intelligences are extremely hard for AI to achieve. Our intelligence is a complex, interwoven whole and it’s here we must focus our attention. We underestimate it at our cost.
“Understanding the difference between human and artificial intelligence will help us make the right decisions about how we use and develop AI.”
HP: What opportunities do you see for AI within education?
RL: One of the most powerful things AI can offer is intelligent analytics – collecting and analysing data to uncover the nuances of how individuals are learning. There’s an increasing appetite within education for learners to develop better metacognitive abilities. Although it’s technically possible to track metacognition with technology, scaled applications are coming but don’t exist yet. Schools are developing processes, approaches and activities to support the development of metacognition – it’s now a question of joining these up with technology.
As an educator, it’s exciting to think that I will be able to understand more about how my students are learning rather than just the output of their learning – educators will be able to understand their students’ needs and provide more targeted support.
HP: What do teachers need to know about AI?
RL: Teachers have no choice but to begin to engage with AI. With the large amount of money invested by technology companies, it will transform education. My key advice is to learn fast but act slowly. There is so much pressure to be pushed into using AI, but now is not the time to rush using some piece of AI without knowing what you want to achieve.
“Teachers have no choice but to begin to engage with AI.”
If you are a school leader, I would immediately get an AI policy in place so that staff and students in your institution understand what is acceptable. This is a chance to pause. Until you know what the risks are, with uploading personal data for instance, and how to protect yourselves from them, it’s not a good idea to use these tools.
HP: How can educators ensure that AI is developed to meet their needs?
RL: We shouldn’t be under any illusion – for the tech companies this is all about money. AI might do great stuff for humanity along the way or it might not. Therefore it’s important to help educators to have meaningful conversations with developers. We are forming a coalition of school and college leaders in the UK to help leaders stay up to date with what is happening in AI in education. We hope to expand internationally in the future. Coalition members will share their learning as they develop an evidence base about what works and what doesn’t when using AI in their organisations. It is a practically focused initiative. A key purpose is to give educators a bigger voice as a group to say what we would like from AI.
We have to make an explicit effort to communicate – it will be good for everyone if tech companies develop what people need and want, and that’s not really happening at the moment. No matter how well technology or any other intervention is designed, it only works well if it’s implemented properly. When it comes to AI, this is probably about having clear objectives that connect to educational goals.
HP: What other opportunities are offered by AI?
RL: We should see AI as an assistant for both teachers and learners. Much of the real power comes when you use AI to connect different activities. For instance, an AI-augmented system can support the production of curriculum content and activities, and also mark and give feedback on these activities when completed by students. Such a system can analyse students’ completion of activities and use AI to suggest improvement to the curriculum and assessment design, so that you are constantly refining. We need to see AI as something that can empower other technologies, to create intelligent virtual and augmented reality, for instance. Obviously you need human oversight.
Two key characteristics of AI are autonomy and adaptivity. Each individual interaction with AI can be powerful in targeting support where a learner needs it. But data about an individual student will give a teacher detailed insights as to where their students are in their learning, and that impacts how a teacher interacts with them. There are so many opportunities, which is why it is so important to mitigate risks and have necessary oversight and safeguards.
Footnotes
Rose Luckin is Professor Emerita at University College London and founder of London-based consultancy Educate Ventures Research (EVR). Rose regularly gives evidence to UK and European policymakers and advises industry at home and abroad.
Her 2022 book AI for School Teachers offers practical, accessible advice for educators. She also wrote the influential Machine Learning and Human Intelligence: The Future of Education for the 21st Century in 2018, and is host of the EdTech Podcast.
Rose has been recognised as a strategic thinker and innovator, and honoured with international awards, including the 2023 ISTE (International Society for Technology in Education) Impact Award. She also sits on a range of advisory boards within education and training, and has taught in schools, universities and within further education.
EVR helps schools, educational institutions and businesses to make evidence-informed use of AI.
Interviewee:
Interview by: