Beyond Standardised Tests — New Goals in Education: Learner Agency supported by AI pedagogy

AI isn’t disrupting education. It’s disrupting the industrial model of education. What’s next? What will post-industrial education look and feel like?

Now that ChatGPT can outperform humans on the robotic tasks we previously used to test young people, it’s prompted many to ask the big questions.

Why are we doing all this education in the first place?

How do we help students understand their unique talents and human qualities that distinguish them from machines?

And how do we redesign schools and our practice to equip them for the modern workplace and yet un-imagined jobs of the future?

In this essay, we’ll explore a new goal for the education system, something fundamental to our human nature.

A goal that is vital to nurture in young people across all contexts — past, present, and especially in our uncertain future.

That goal is to develop Learner Agency.

The future of skills is a moving target. Any structures that don’t enable students to explore and understand their fundamental human nature will be built on shifting sand. Whatever the future holds it’s our ability to adapt and respond which is central to supporting the next generation to thrive through uncertain times.

“…education must teach the individual how to classify and reclassify information, how to evaluate its veracity, how to change categories when necessary, how to move from the concrete to the abstract and back, how to look at problems from a new direction” — Herbert Gerjuoy

In later essays and videos, we’ll explore how you can robustly measure Learner Agency based on our work at WILD Learning Sciences.

But for now, let’s delve into some definitions and conclude with some strategies you can apply in the classroom to support your learners.

What is Agency?

Agency is a fundamental part of the human condition.

It’s the core concept underlying terms such as:

• Self-directed learning
• Autonomous learning
• Self-determined mindset
• Self-leadership
• Intelligence

The common definition of agency is “the capacity to take action”. However, this is overly simplistic. It would suggest that even a thermostat could be seen as having agency because it reacts to changes in temperature.

To construct a notion of agency, we need to understand and integrate ideas that enable a simple controller (like a thermostat) to plan and exhibit goal-oriented behavior.

From Thermostats to Agency

Consider the thermostat as a basic controller. It operates on a simple feedback loop: when the temperature drops below a certain set point, the thermostat “reacts” by turning the heating system on. When the temperature rises above the set point, it turns the system off.

This is reactive control, not agency.

Agency requires an additional step: a system must predict and evaluate future states by forming counterfactuals or many “what-if” scenarios.

Why? Because the future universe doesn’t exist right now, you need to create a counterfactual universe.

This introduces a branching universe of possibilities where agents take action towards the most desirable outcome.

Okay now we’ve added the ability to form counterfactuals to our controller.

Counterfactuals are the minimum requirement for a system to exhibit agency. Thus, agency can be more accurately defined as:

The capacity of a system to form counterfactuals, create and use internal models of itself and the world for prediction, and act purposefully to optimize desired future states towards goals.

To construct the idea of agency fully we’d need to speak to the role of motivation, beliefs, identity, rewards and evolutionary selection pressure and so on. For brevity, let’s press on!

Depending on the kind of agent — be it a bug, a virus, or a human — the goal-directed behavior will look different.

Because humans are a certain kind of agent, with arms, organs and brains of a certain size, we exhibit different behaviour towards our goals.

We have biologically rooted cognitive processes and social-emotional capacities which are essential for us to make sense of an uncertain world, learn, form models that allow us to use information and our environments to change the universe in front of us towards our desired state.

We create beliefs, experience social lives, drive cars, and re-arrange information around ourselves to solve problems so we can pursue our goals.

If these cognitive processes and social emotional capacities are stifled in some way, our agency collapses.

Here’s just a few ways schools collapse agency.

With our high level definition of agency it shouldn’t be hard to see that the opposite of having agency would involve:

a) Preventing learners from having the ability to physically move to achieve their goals

b) Preventing learners from having ideas about their future and create plans for future states

c) Preventing learners from understanding how to regulate their emotions to achieve their goals

d) Preventing learners from making new social relationships and collaborating to build capacity, knowledge, conviction and support to achieve their goals.

How to support agency as an Augmented Educator

The support that humans need to develop agency towards their goals is different from what flies, whales, or robots require. These entities have different brain and body structures, and form different goals.

This is our task as educators

• We need a shared language to discuss what supports humans to develop agency, and to engage in conversations with our learners about their own agency.
• But more than just talking about it, we need to be capable of building networks of relationships, curating and critiquing information sources, and creating interventions around a learner to increase their agency.

To do this, we’ll utilize the language of Learning Power.

The concept of Learning Power was popularized in the 1980s and 90s by cognitive scientist Guy Claxton. It was later developed into a robust assessment tool and language by Ruth Crick and the team at WILD Learning Sciences.

Learning Power is a set of eight holistic dispositions that enable a human to navigate and make sense of the world as they approach risk and uncertainty in pursuit of their goals.

Let’s examine these dimensions, pausing at each to envision a scenario where a learner is lacking in each dimension.

Mindful Agency:

This involves seeing the big picture, taking responsibility, being proactive, and managing emotions. The opposite is being ‘robotic’.

If a learner lacks mindful agency, they’re unable to define a goal, fail to see the point of the goal, and struggle to regulate their emotions towards it.

Hope & Optimism:

This means seeing oneself as a learner who grows and improves over time, possessing a growth mindset. The opposite is being ‘stuck and static’.

When hope and optimism is low, a learner can become defeatist, and their self-talk emphasizes how ‘stuck’ they are or how insurmountable a task seems.

Sense-making:

This involves seeing connections, making meaning, and relating new learning to what matters to the learner. The opposite is experiencing fragmentation and confusion.

If a learner can’t make sense of information, they might disengage from a group due to confusion, take a wrong path, or seek new information sources.

Creativity:

This includes risk-taking, playfulness, and using imagination and intuition in learning. The opposite is being ‘rule-bound’.

When learners lack creativity, they struggle to create novel variations of information or objects. They might resort to existing strategies for solving a problem or struggle to imagine alternative methods.

Curiosity:

This involves asking questions, desiring to learn more, and delving below the surface. The opposite is taking things at face value.

If learners lack curiosity, they passively receive information. Their questions don’t probe to understand concepts from multiple perspectives. They stop asking ‘how’, ‘why’, ‘when’, and ‘who’.

Collaboration:

This involves interacting with fellow learners, providing and receiving support, and solving problems through teamwork. The opposite is being isolated.

If learners struggle with collaboration, they might carry too much burden without asking for help, fail to synthesize diverse ideas in a group, or evaluate diverse perspectives.

Belonging:

This is feeling supported, encouraged, and at ease in a learning community. The opposite is feeling ‘out of place’.

There are many indicators if learners don’t feel a sense of belonging with peers, mentors, teachers, or experts. Some examples include imposter syndrome and a lack of safety.

Orientation to learning:

This means being ready and open to learning, striking a balance between determination and adaptability. It involves not being too fragile and dependent, nor too rigid and persistent. Just right of centre, a bit towards persistence is optimal.

If a learner is rigidly persistent, they believe that their way is the only way and can’t change tactics. They will be closed-minded to new ideas or approaches.

Conversely, if a learner is too ‘fragile dependent’ they are unable to work independently, doing everything because someone else directs them or provides the answer.

Learning Power Underpins All Human Goals

Can you recall a time when you set out to pursue a goal, perhaps a promotion, a move to a new country, or a career change? Can you identify which dimensions of learning power you might have been lacking in?

We can even use learning power assessments to identify ‘at-risk’ or disengaged learners.

This concept of learning power is central to how any human approaches a goal, whether it’s getting good grades, being a headmaster, managing a company, or learning to dance.

If appropriately supported, individuals can gain agency and become adaptive, self-directed learners who take ownership of their growth.

They become the kind of self-determined person that people want to hire for management and leadership positions.

Most importantly they know what to do when they don’t know what to do – and become the driver of their own learning.

Put it into action — Listen to self talk

Next time your children, colleagues, friends or learners talk about their work or goals, listen to their self talk and consider if they’re depressed in any of these dispositions.

Practise initiating conversations using these terms with your children, colleagues, friends, and learners. Ask them if they’re struggling to come up with creative solutions to a problem, or if they’re lacking belonging with peers or mentors in their new project, or if they’re struggling to make sense of new information.

If they confirm that they are, then ask them:“What would support you to think of new ideas or approaches, feel a stronger sense of belonging or make sense?”

Take the position of a coach and support them in making their own decision.

If they need help, brainstorm strategies to find, create or ask for that support.

The ultimate goal is to enable them to identify what would help them based on their own assessment of which learning power dimension they’re lacking in.

Placing learning power in the classroom

Okay so you’ve been introduced to the core language of learning power, how you might begin to support learners, started to develop a sensitivity for how schools stifle it and how central it is for human flourishing.

On top of just knowing the language there are a few more capacities that the new breed of educator will need in times of rapid change (And even not in times of rapid change). The’ll need the capacity to…

1. Understand and talk about their own agency in terms of learning power and then be able to support learning power in others.
2. Build a web of learning relationships around a learner to support their goals.
3. Develop a learner’s capability for systems thinking and complex problem solving.

I will get into how educators can build these capacities in more detail in later essays.

Depending on how much you lean into the above 3 capabilities and incorporate them into your lesson plan, there are big and small ways your classes will change.

The tempo of classes, who and what is delivering content and instruction, where knowledge is created and stored might all be very different.

When this kind of pedagogy is central as it is in some progressive schools and student led environments, the physical classroom begins to look very different. They become redesigned to better enable this way of learning.

It’s no coincidence that these classrooms mimic modern multifunctional workspaces, co-working spaces, and spaceship factory floors, as these are areas where adults already utilize and refine the aforementioned three capabilities in their teams.

Where classroom practice doesn’t adopt this way of learning, spaces will remain frozen in time, modeled after the workplaces of our industrial past.

A call for new classroom praxis for Learning Power.

The need for a new praxis is urgent. Many curriculums take 10 years to go from idea to implementation. At our rate of change. This simply wont keep pace.

I don’t mean to suggest that we must all rapidly adopt state-of- the art curriculum on emerging technology and global challenges — though I encourage it — but the way we teach science and silo the traditional subjects will cause us to build a fragile, unprepared society.

Teachers are the front of innovation now. Not the software. We have all the tools but need culture change. Our collective ability to model these new ‘hows’ of learning will be the gateway of how young people understand how to re apply those methods to any new skill set or information.

Let’s build a new praxis with learner agency and AI at its core

📺 Here’s a demo video

Teachers are not being replaced by AI, their role will just look substantially different.

We’ll need to create new instruction and activities that leverage just-in-time knowledge, peer learning, systems mapping and regular formal and informal coaching checkins with each learner.

This will be especially true for middle school upwards where learning shifts from literacy and numeracy fundamentals towards giving teenagers basic theories of knowledge, the space to develop their identities, professional networks and vocational content.

This new praxis has many new features. But let’s get into one characteristic aspect.

ChatGPT prompts for supporting each dimension of learning power

Now that you’re getting good at listening and guiding your learner to identify which learning power dimension they’re low in you’ll want to give yourself and them some scaffolding or a first step for coming up with some strategies for supporting them.

If your learners understand how to use these prompts they can become more autonomous over time. When you’re adept in identifying learning power you’ll be able to create support with them on the fly without chatGPT.

Mindful Agency:

• I’m working on ____. Help me to list potential obstacles I might face and strategies to overcome them.
• I’m studying ____. Suggest a set of achievable goals and milestones for my project/study.
• I’m trying to achieve ____. Provide me with resources on how to manage my emotions effectively during this process.

Hope & Optimism:

• I’m learning ____. Share inspiring stories or case studies of people who have successfully mastered this topic.
• I’m studying ____. Suggest some positive affirmations or mantras I can use when I feel stuck or discouraged.
• I’m working towards ____. Help me to identify smaller, achievable goals that can lead to success in this area.

Sense-making:

• I’m studying ____. List the key concepts in a structured order, starting with the most fundamental and progressing to the more complex, with an explanation of how each builds on the previous.
• I’m learning about ____. Suggest some practical exercises or examples that can help me apply these concepts to real-world scenarios.
• I’m trying to understand ____. Show me some diagrams, infographics or videos that could help clarify this concept.

Creativity:

• I’m working on ____. Suggest some brainstorming techniques or creative exercises that could help me think outside the box.
• I’m studying ____. Share some innovative approaches or strategies used in this field.
• I’m trying to create ____. Help me find some inspirational sources or examples from related fields that could stimulate my creativity.

Curiosity:

• I’m learning about ____. What are some deeper, intriguing questions I should be asking about this topic?
• I’m studying ____. Suggest some additional resources or reading material to help me explore this topic further.
• I’m exploring ____. Show me some current debates or unsolved mysteries in this field.

Collaboration:

• I’m working on ____. Recommend some online communities or forums where I can interact with fellow learners.
• I’m studying ____. Help me find potential study partners or mentors in this field.
• I’m engaged in ____. Show me some tools or platforms I can use to collaborate with others online.

Belonging:

• I’m studying ____. Compile a list of relevant experts in this field, along with their professional profiles and affiliations.
• I’m learning about ____. Recommend some online groups, clubs, or communities where I can connect with like-minded learners.
• I’m interested in ____. Find me some popular blogs, podcasts, or YouTube channels related to this topic.

Orientation to Learning (Persistence):

• I’m studying ____. Suggest some strategies for dealing with frustration or setbacks in this field.
• I’m working on ____. The goals is to ____ My current approach is ____ Share alternative approaches to the same goal.
• I’m learning ____. My Recommend some resources or exercises to help me develop my resilience and tenacity.

Orientation to Learning (Dependence):

• I’m learning about ____. My goal is ____ Suggest the next three steps I should take which utlaise independent study techniques or strategies.
• I’m studying ____. Show me how to effectively use resources (books, internet, etc.) for self-learning in this field.
• I’m working on ____. Help me create a self-paced learning plan with goals, tasks, and deadlines.

I’ve sketched these quickly with the help of chatGPT. They’re not perfect but hopefully you can adapt them for your context and create other iterations of them amongst your peers. Here’s a transcript of the conversation with chatGPT so you can recreate them.

The main point is that you can be aware of the fronts that AI can augment human creativity. This should feel significantly different to preparing them for grades. With this kind of foundation and an alignment of curriculum with a learner’s authentic / intrinsic motivations our education could support a very different kind of sovereign an self-determined young person.

If we can be intentional in our praxis design, this intersection of autodidactic learning & AI will be a huge stepping stone towards unleashing creativity, unlocking productivity and upleveling skills.

Done with care, rigor and a willingness to let the learner take lead, we will be well be on our way to a golden age of pedagogy and human ingenuity.

Wrap up: New Goals in Education — New Goals for Our Learners?

Throughout this essay, it’s been left unsaid what the learners’ goals are, why their goals are what they are, and who should have the teacherly authority to give or revoke those goals in the first place.

Schools and universities have been going through a legitimation crisis for decades now. ChatGPT has just brought that to the fore.

It’s time we take a hard look at the middle swather of ‘filler’ curriculum, degrees and silod humanities topics and redesign them to be interdisciplinary, applied, personalised or even student driven. All of this is within reach now with AI.

“Over the next century, scholars and fans aided by computational algorithms, will knit together the books of the world into a single networked literature. A reader will be able to generate a social graph of an idea, or timeline of a concept, or a networked map of influence for any notion in the library. We’ll come to understand that no work, no idea, stands truly alone, but that all good true and beautiful things are networks, ecosystems of intertwined parts, related entities and similar works” — Kevin Kelly — Editor Wired Magazine

Learners aren’t dumb, unless they’re in ivy league or a progressive school they complain about half the content we assign them is irrelevant and a waste of their time. If they don’t it’s just because we’ve drilled compliance into them with a bell and blackboard.

We assign them goals set by the curriculum. But does the learner authentically desire this goal? How might the curriculum become personalised to what motivates them.

Are all the parts of the curriculum worth sequencing and assessing? Literacy, Numeracy, sure, but might history and geography and a second languages be delivered in a ‘just in time’ way — delivered when the learner’s current goals activate an intrinsic drive to learn about a specific and related history, geography or language?

When curriculum is out of date and the world outside of the classroom looks different to what’s being taught inside the classroom how will we make the curriculum relevant quickly?

When is a learner ready to take responsibility of their learning and destiny, to shape their environment and act on opportunities?

Is it determined by some arbitrary age? When they’re ‘mature’ enough? When they hold a certain credential? When they’re self-determined enough?

How would you even know when they’re ready?

In the next essay i’ll get into that.

For now I want to leave you with this.

A future of work to prepare for (Incase you didn’t think change was coming quick enough)

In a report by MBO and in similar reports from OECD around 60% of the American workforce will be independant or freelance workers as work becomes more ‘fractionalised’ —perhaps performing many job functions throughout the day at multiple different organisations. A marketer in the morning, A child play space facilitator in the middle of the day and a Software developer in the evening.

Source

In our near future, living, learning, and work will be integrated, changing when, where and how ‘education’ is delivered.

We need to use the best pre-made ‘recipes’ for learning numeracy and literacy rapidly from Maria Montessori, melded with the playful immersion of systems thinking of Seymour Papert combined with the dynamism of self-directed learning afforded by AI and Ruth Crick.

Perhaps we could call this new approach ‘Work Integrated Learning Design’ or WILD learning 😉

As ‘work’, both in the Montessorian sense, and in the future of integrated work sense are central to getting future generations ready to thrive and not just survive.

My team at City as a School are building infrastructure to mainstream this way of learning. We’re in the early stages with initial traction and would appreciate partners for the journey.

For educators we currently train cohorts of educators to do this at WILD. I’d love to meet you at one of the trainings soon.

With care from London

Serj and the CAS team.

Underpinning research

The following papers are representative of a much larger body of work into Learning Power led by Ruth Crick. See their bibliographies to go deeper.

Crick, R., and Bentley, J. (2020). Becoming a Resilient Organisation: Integrating People and Practice in Infrastructure Services. International Journal of Sustainable Engineering, pp.1–18. Published Online: 1 June, 2020. DOI: https://doi.org/10.1080/19397038.2020.1750738. [Staff development case study connecting Learning Power to organisational staff survey results using VOICE]

Gardner, A., Goldfinch, T. and Willey, K. (2017). Characterising the learning dispositions of first year engineering students. Proceedings AAEE2017: Australian Assoc. for Engineering Education, Sydney. [PDF] [Learning Resources] [Quant/qual analysis of student dispositions, and a learning design for engaging students with Learning Power]

Barratt-See, G., Cheng, M., Deakin Crick, R. & Buckingham Shum, S. (2017). Assessing Resilient Agency with CLARA: Empirical Findings from Piloting a Visual Analytics Tool at UTS. Proceedings UniSTARS 2017: University Students, Transitions, Achievement, Retention & Success. (Adelaide, 1–4 July, 2017). [PDF] [Summary of a large internal UTS analysis]

Deakin Crick, R., S. Huang, A. Ahmed-Shafi and C. Goldspink (2015). Developing Resilient Agency in Learning: The Internal Structure of Learning Power. British Journal of Educational Studies 63(2): 121- 160. DOI and Open Access Eprint: http://dx.doi.org/10.1080/00071005.2015.1006574 [Psychometric basis for the construct of Resilient Agency underpinning the self-assessment survey]

Deakin Crick, R. & Goldspink, C. (2014): Learner Dispositions, Self-Theories and Student Engagement. British Journal of Educational Studies, 62(1), pp.19–35. DOI: http://dx.doi.org/10.1080/00071005.2014.904038. Open Access Eprint: PDF [Investigates different conceptions of learning dispositions]

Buckingham Shum, S. and Deakin Crick, R. (2012). Learning Dispositions and Transferable Competencies: Pedagogy, Modelling and Learning Analytics. Proc. 2nd Int. Conf. Learning Analytics & Knowledge. (29 Apr-2 May, 2012, Vancouver, BC). ACM Press: New York. DOI: 10.1145/2330601.2330629. Open Access Eprint: http://oro.open.ac.uk/32823 (conference replay) [Established Dispositional Learning Analytics as a field]

Learning Power