What Human Capabilities are we Cultivating?

While we bring an increasing range of frameworks and technologies to measuring what people can do, we increasingly lose sight of the capabilities we are trying to grow in them, and our systems haven’t noticed.

I have spent a large proportion of my working life developing measurement tools and evaluating reform initiatives in different aspects of the education system. The question I’m typically asked to assess is whether the project met its objectives – was there improvement in student learning, was teacher quality of practice improved etc etc. The question I’m not asked, is what does ‘good’ look like. What does being a ‘good’ school or University actually mean? What are ‘good’ learning outcomes? This is always assumed and rarely questioned. Evaluation projects increasingly ask us to examine data sets defined by what was easy to measure, not what needed to be measured.

Systems have become very good at measuring attainment while slowly losing sight of what the attainment was ever meant to represent. The implication is that the gap between what the education systems support and what learners need is widening.

Much of the current debate about skills, credentials and the future of work is really a debate between measuring what is easy to measure and measuring what matters (which may not be as easy (cheap) to measure). How do we verify what a person can do, how do we make a record portable and trustworthy, how do we keep it meaningful in a world where a convincing account of capability can be generated at almost no cost. These are serious questions and they deserve serious answers. But they all sit downstream of the more fundamental questions education systems increasingly pass over. Before we ask how to measure capability, it is worth asking which human capabilities are worth cultivating, and whether the things we find easiest to measure are the things that matter most.

The capabilities we rarely measure

Vivienne Ming, argues that the capacities which actually predict how a person will fare across a career are diverse and largely independent of one another, and that the one that matters most is meta-learning: the ability to keep learning as the ground moves beneath you. Meta-learning is not a subject, and not a module to be completed. It is a capacity that forms slowly with the appropriate support, by being asked again and again to understand and improve complex unfamiliar problems. It will not appear through simplistic graduate surveys and I have been told, by people who should know better, that it cannot be measured. Yet interestingly if you asked a group of co-workers to identify which of their colleagues have this capacity they typically have no difficulty identifying them.

So, we find ourselves with an education system drifting further away from the human capabilities that many argue are needed, at least partly due to the measurement approaches that are being applied. When a system is rewarded for what it measures, it tends to measure what is easily measurable, and then to teach toward it. The purpose of the system then becomes a lag indicator rather than a driver. We have watched this happen before, with standardised testing, with the counting of contact hours, with the steady elevation of completion over comprehension. A skills-first era, for all its real merits, can repeat the error on a larger scale if it treats a skill as a discrete, checkable unit and forgets that the most valuable human capacities are integrative and slow to form. Getting better at measurement, without getting clearer about what we are measuring for, is only a way of moving faster in a direction we have not paused to examine.

Cultivation, and what we are defunding

There is an irony in this that educators feel keenly. The part of our tradition that has always concerned itself with supporting the development of these higher order capabilities, the broad and increasingly unfashionable liberal education, is the part now under the most pressure to justify its immediate worth. Yet its whole purpose is the formation of the general capacities, judgement, the holding of competing ideas, the quick entry into an unfamiliar field, that we now insist we most need and can least quantify. 

Joseph Aoun made a version of this argument in his work on higher education and automation, holding that what makes a graduate resistant to displacement is exactly the higher-order, distinctly human capacity that a narrow and vocational reading of education tends to crowd out. We are discounting the cultivation of adaptability at the very moment we are demanding more of it.

None of this is an argument against measuring outcomes. Measuring what a person can actually do, rather than the educational input they received, is critical to employers, and value derived by the learner. The institutions that learn to do it well have the greatest chance of remaining relevant in an AI enabled world.

What I notice is that the institutions thinking most seriously about this, the professional bodies, the universities willing to interrogate their own habits, the assurers prepared to change what they reward, are not approaching it as a problem of method alone. They are treating it as a question about what they are for. That strikes me as the right place to begin, and a more hopeful one, because unless approaches to outcome measurement and the cultivation of capability advance together, the first will continue to hollow out the second.

Why this matters now more than ever

It matters now because Australia is redesigning the rules of capability recognition while barely touching the question of capability itself. Jobs and Skills Australia is publishing frameworks that will shape how learning is defined; Department of Employment and Workplace Relations' Guiding Principles for Assessing Authorities and the National Skills Passport are building the pipes through which those definitions will flow; migration reform is rewriting who gets to be recognised as “skilled” in the first place. 

These are structural moves, not marginal ones. If we don’t decide which human capacities these systems are meant to protect, they will default to what is easiest to codify and reward. In a reform wave this large, the risk isn’t that we measure the wrong things – it’s that we build national infrastructure that makes it harder to measure the right ones. That’s why the question of what we are cultivating cannot wait.

---

Reading

• Vivienne Ming, on meta-learning and the diverse predictors of long-run capability.

• Joseph Aoun, Robot-Proof: Higher Education in the Age of Artificial Intelligence (MIT Press, 2017; revised edition 2024).