On the last mile problem in education

Edtech desperately needs to solve for human motivation

For someone who is critical of edtech, I am pretty tech-obsessed in my personal life, particularly in the area of fitness and longevity. I own an Apple Watch and a Whoop, and I use ChatGPT for all sorts of wellness hacks. I went through a phase of using ChatGPT like MyFitnessPal, and I make my own preworkout from a ChatGPT recipe that I am assured will not accidentally kill me.

But fitness is expensive and along with my obsessions, my watch face must match my outfit each day, and my Whoop band along with it. Those little nylon bands (yes, that’s all they are) cost upwards of about $120, so imagine my delight when I discovered those bands on Shein for $8 each. The bands in brown and navy (unavailable colours from Whoop) were shipped from China with mind-blowing efficiency. Within about 48 hours they were sitting in a warehouse not that far from my house.

And then I waited a week and a half. This last mile problem is a known issue in logistics, telecommunications, healthcare and public transport. It’s a metaphor for the disproportionate cost, complexity, and inefficiency found in the final step of any network. And I think it’s a problem in education too.

It was big news a couple of months ago when Sal Khan’s chatbot was quietly killed off. For those of you who don’t know, Khan Academy started as a series of online videos, much like Mister Woo-Tube, but evolved to be more interactive over time, giving clues instead of solutions. I had kids in primary school at the beginning of the site’s evolution. My own girls found it helpful but a bit limited compared to having a human just prompt or cue clues with the Goldilocks level of support¹.

Khan Academy was a magnet for funding. Given the track record of relative success compared to other edtech endeavours, Khan should have been able to make it work. It seems that the issue wasn’t the tool, but the challenge of getting students to use it. In this case, human motivation is the last mile – the distance from whatever the kid would rather be doing, to sitting in the seat in front of the screen.

I didn’t initially know the extent to which this problem was perennial. I recently heard the eminently wise Daisy Christodoulou talking about AI on Craig Barton’s podcast, where she explained that non-human tutors and personalised learning has been around since the 90s. The issue was never, and has never, been the limitations of tech, which has the potential to be as great as the best available human inputs. However, it may be the case that as much as students moan about school, they may actually like their human tutors and socialised learning. I’m not a lover of the word “motivate” as a verb, because even teenagers are naturally inclined towards growth. But it’s the actions of teachers that provide the quality of the soil and nutrients for that growth. As yet, edtech is not up to this job.

I predict similar issues with models like Alpha School. Let’s ignore the fact for now that the school is pegged at an elite price point of around $40-60K USD per year, and that advantaged students will come with more than a modicum of background knowledge. The model (as I understand it) employs a technology-enabled session of Direct Instruction, overseen by facilitators instead of trained teachers. Students have the afternoon free for passion projects. Their learning in basic skills is accelerated by the responsiveness of the tech, and students appear to be highly extrinsically motivated to finish their work and unlock the reward of afternoon (relative) freedom.

The learning gains are impressive, but the model itself appears highly transactional, in motivation terms. Learning (a race, an obligation) is a chore to be completed in exchange for doing what students really care about. I’m an English teacher, so it’s hard for me to imagine how young people might be motivated to learn about war poetry or Shakespeare under this model that seems to sideline a knowledge-rich curriculum.

In my early career, I was part of a cross-curricular team that programmed similar time for such passion projects, and it seemed to me that the scope and quality of those projects were severely constrained by the students’ prior achievement level. One gifted student took about six hours of learning time to develop a glove for non-verbal people with a disability, which was coded using Scratch to give voice to basic commands, using only touch. Another took that same six hours to develop a new flavour of teabag. It does seem that the Alpha model has solved the last mile problem in one sense, but perhaps at the expense of more cognitively demanding knowledge and skills.

The problem is common in coaching programs too. The observations happen, the cups of coffee and nice chats happen, but then little change is seen in the classroom. Having a coaching program is a prestige-piece in many schools. Having a stagnant coaching program is far more common and is the quiet part that nobody says out loud.

Several online or AI coaching models are hitting the market, and I can see this being part of our current edtech/AI goldrush. These models can provide brilliant feedback, but again, the challenge is getting the teacher to film, review and act, based on the advice of a non-human coach. The advantage of AI coaching is that these models are potentially scalable in situations where it’s difficult and expensive to release teachers from class. But if human coaches struggle, online coaches have an even bigger obstacle – persuading the coachee to spend time with the feedback and act on it without the healthy social pressure of complying with another human’s expectations.

I used to think that concerns about motivation were a bit of a bypass for educators who didn’t want to talk about quality teaching and curriculum, or who wanted to shift blame to students when everything was seemingly done right but nothing was really working. I’m not saying by any means that I have solved the problem: it’s a wicked one and will only intensify as edtech muscles in on teachers in greater numbers and students increasingly see school as irrelevant and optional.

The last mile in logistics is traversed by a low-paid human in a van from a warehouse in Chullora. And as much as I love my wearable tech, my true motivator is my fitfam, the half-asleep, half-coherent hooligans who get up at 430am most mornings to lift heavy things and put them down again, all to the tune of music I am way too old for.

In the same way, the last mile in education is from coach to teacher or teacher to student. And I feel pretty confident that both parties will need to be human for the foreseeable future.

1. Side note: I ran a workshop on academic wellbeing recently and pre-empted questions about AI tools standing in for spaced practice, interleaving and retrieval. It turns out these tools were not on the radar for the participants, perhaps due to the idea that getting students to sit diligently at a laptop when there are other far more interesting things on screens was not realistic.