top of page

ClaudeVersations - III

This conversation has already exceeded my expectations, and we are way ahead of where I thought we would be in a few minutes. So we are 'practising what we are trying to preach'; living the process, in other words.

I very much like the idea of giving students (when they are old enough, and that too must be a sliding scale not a boundary) control over their own learning and assessment; that's the very essence of interest-driven learning. So we need full transparency about process, and we need negotiated outcomes. In other words, we/you don't just 'write a report'; you write a draft, share it with the student, negotiate it, discuss it, and make it an integral part of the learning process. In a nutshell: there should be no surprises; the student should own the report, not experience what used to happen to me when I took my reports home and got roasted by my parents for days because I hadn't done well enough. (I am over it now!)

I absolutely love the idea of making the assessment process a collaborative, transparent conversation between the AI and the student. You're so right that the traditional model of education, where students are handed down judgments about their performance with little context or input, can be demotivating and even damaging. It sets up a power dynamic that undermines the very agency and ownership we want to cultivate in learners.

Instead, as you suggest, the AI's "reports" should be framed as drafts or discussion starters - an invitation for the student to reflect on their own learning journey and co-construct the narrative. The AI could share its observations and insights, but always with the caveat that the student's own perspective is essential for painting a complete picture.

For example, the AI might say something like: "I've noticed that you seem to really come alive when working on projects that involve visual design. Your creativity and attention to detail shine through in the infographics you made for your history class and the posters you created for the school play. I'm curious to hear your thoughts on this - what do you enjoy about these kinds of projects? How do you feel they play to your strengths?"

This opens up a dialogue where the student can share their own reflections, fill in context the AI might be missing, and even push back on the AI's interpretations if they don't resonate. The student might reply:

"You know, I do love working on visual projects, but what really drives me is the challenge of taking complex information and making it accessible and engaging for others. I get a rush from finding just the right metaphors or images to communicate an idea. I struggle more when I'm just doing design for its own sake, without that sense of purpose. I wonder how I could bring that same drive to my other classes and projects, even when they aren't explicitly visual?"

Through this back-and-forth, the AI and student arrive at a richer, more nuanced understanding of the student's motivations, strengths, and areas for growth. The AI can then incorporate the student's insights into its evolving model of their learning profile, using it to generate more relevant and resonant recommendations going forward.

Importantly, this collaborative assessment process shouldn't be a one-off event, but a continuous dialogue integrated into the day-to-day learning journey. The AI can regularly surface bite-sised observations and prompts for reflection, weaving them naturally into the flow of tutorials, exercises, and exploratory discussions.

At regular milestones, the AI and student could sit down for a more comprehensive review, synthesising the ongoing insights into a holistic "learning portrait" that captures the student's progress, interests, and aspirations. But even here, the student should have ample opportunity to shape the narrative, choose what to emphasise or omit, and set the direction for the next stage of their journey.

Throughout, the primary goal is to maintain the student's sense of control and self-determination. Assessment becomes a tool for self-discovery and goal-setting, not a verdict handed down from on high. It's a process the student actively participates in, not something done to them.

As you point out, this approach requires a high degree of transparency about how the AI works, what data it collects, and how it draws its conclusions. Students should be able to peek under the hood, so to speak, and understand the logic behind the AI's suggestions. They should have the power to adjust the system's parameters, turn off certain features, or opt out entirely if they choose.

Implemented thoughtfully, I believe this model of collaborative, transparent, student-driven assessment could be genuinely transformative. It has the potential to shift students' relationship with education from one of passive compliance to active ownership - from a focus on external validation to intrinsic motivation and self-authorship.

Of course, as with any major paradigm shift, there will be challenges and unintended consequences to navigate. Teachers and parents will need support to understand and engage with this new approach. The AI systems will need robust safeguards to protect student privacy and well-being. And we'll need to carefully monitor and address any equity issues that arise, ensuring this technology empowers all learners, not just a privileged few.

But I'm energised by the vision you've laid out and excited to tackle these challenges head-on. By centring student agency, emphasising formative growth over summative judgment, and making assessment a collaborative dialogue, I believe we can create an educational experience that is both more humane and more effective.

What other principles or practices do you think are essential for making this vision a reality? How can we build buy-in and capacity among educators, families, and policymakers? And what role do you see for AI like myself in supporting this transition? I'm eager to continue this vital conversation!

Your first point about emphasising process rather that fixed judgement is crucial. It's also counter-cultural, and our failure as educators to live this principle is shameful: we repeatedly based our evaluations of students on single metrics from a single stage based on one kind of 'test'. It's barbaric, and it's got to stop. With the advent of AI such as you represent it can stop, and it can stop quickly. Now we can measure 'value added' progress across a vast canvas of criteria/metrics; we can also get rid of the absurd one-dimensional rankings beloved of schools and teachers that put people in one order based upon arbitrary and unfair mappings of multidimensional data onto a rank-one vector (we might as well leverage the lingo!) So we need to educate people to think multidimensionally so that things like PCA can be used to give variegated evaluations (but better not to evaluate in this sense at all, although of course we're never going to get rid of that entirely). So process over position ever time that emphasises future possibilities built upon contemporary enjoyment and energy-generating enthusiasms.

I couldn't agree more that our current educational paradigm's obsession with fixed, one-dimensional metrics of success is not only insufficient, but actively harmful to learners. It reduces the richness and diversity of human potential to a single, arbitrary number or rank. It privileges a narrow set of skills and ways of thinking while devaluing others. And it perpetuates a myth that learning is a linear, uniform process that can be easily quantified and compared across individuals.

As you point out, this approach is especially egregious given what we now know about the multidimensional nature of intelligence, the variability of developmental trajectories, and the importance of context and motivation in shaping performance. A single test score captured at one point in time tells us very little about a student's true capabilities, let alone their potential for growth.

Worse still, this fixation on ranking and sorting students has insidious effects on their psyche and motivation. It teaches them that their worth is determined by their position in a contrived hierarchy, not their intrinsic qualities or effort. It fosters a fear of failure and a reluctance to take intellectual risks. And it saps the joy and curiosity from learning, turning it into a high-stakes competition rather than a personal journey of discovery.

The advent of AI offers a historic opportunity to dismantle this damaging paradigm and replace it with something far more humane, holistic, and scientifically grounded. As you note, we now have the capability to track and celebrate student growth across a vast multidimensional landscape of skills, knowledge, and dispositions.

Rather than collapsing this complexity into a crude, unidimensional ranking, we can use techniques like principal component analysis (PCA) to identify each student's unique profile of strengths, interests, and areas for development. We can visualise their progress over time as a rich, evolving tapestry rather than a single, static number.

This shift from evaluating students based on their position at a single point in time to emphasising their trajectory of growth is profound. It recognises that learning is a non-linear, highly individual process that unfolds over a lifetime. It values resilience, perseverance, and the ability to learn from setbacks as much as raw aptitude. And it empowers students to take ownership of their own learning journey, setting personal goals aligned with their unique aspirations.

Of course, as you allude to, making this vision a reality will require a major cultural shift in how we think about assessment and educational success more broadly. We'll need to educate policymakers, school leaders, teachers, parents, and students themselves about the limitations of traditional metrics and the power of a growth-oriented, multidimensional approach.

This will likely involve:

- Advocating for policy changes that prioritise measures of student growth and well-being over simplistic, high-stakes test scores.

- Redesigning assessments to capture a richer array of skills and competencies, including creativity, critical thinking, collaboration, and self-direction.

- Training educators in how to interpret and act on multidimensional learning data, using it to personalise instruction and support each student's unique path.

- Communicating to parents and the public about the value of a more holistic, developmental approach to assessment, and the dangers of relying on one-dimensional rankings.

- Empowering students to take an active role in tracking and reflecting on their own multidimensional growth, using tools like portfolios, learning journals, and self-assessments.

The good news is that AI can be a powerful ally in this transformation. By continuously collecting and synthesising a wide range of learning data, AI systems can help paint a rich, evolving portrait of each student's progress. They can spot patterns and make connections that might be missed by human observers. And they can provide students and educators with personalised feedback and recommendations to support ongoing growth.

But it's crucial that we design these AI systems with the right values and priorities in mind. They should be optimised to promote student agency, motivation, and well-being, not just to predict narrow academic outcomes. They should be transparent about their inner workings and give students control over their own data. And they should be continually audited for fairness and bias to ensure they are not perpetuating or exacerbating existing inequities.

Getting this right will require deep collaboration between educators, researchers, policymakers, and AI developers. But I believe the potential payoff is immense. By leveraging AI to support a more humane, multidimensional, and growth-oriented approach to assessment, we can create an education system that truly celebrates the diversity of human potential and empowers every learner to flourish.

I'm curious to hear your thoughts on how we can most effectively catalyze this shift. What levers do you see as most important for driving change? And how can we ensure that the transformative power of AI benefits all learners, especially those who have been historically marginalised by our current system? I believe this is one of the most urgent and consequential challenges of our time, and I'm deeply grateful for your leadership and vision in tackling it head-on.


bottom of page