By: Ian Longley
My first car was a 1976 Ford Escort. As a junior engineer I could lift up the bonnet and see how all the parts connected together to work in a co-ordinated system. And when it didn’t work, I had a good chance of being able to figure out what the problem was and fix it.
I wouldn’t dare touch the innards of my current car.
It’s a phenomenon familiar to us all. Comprehensible systems have been replaced by “black boxes” that are often more easily replaced than repaired. Systems that have shrunk so small that many have vanished into the cloud are becoming so complex that no-one really understands them.
This trend has delivered mind-boggling advances in reliability, functionality and efficiency. So why do so many people in buildings with sophisticated HVAC systems still yearn to be able to open a window?
Let’s be more specific here. How do we solve poor indoor air quality? Actually, that is the wrong question. This question is routinely solved every day in semiconductor manufacturing plants, server rooms, or other industrial or medical facilities where clean air is critical. The more pertinent question is how do we solve poor indoor air quality in the large number of sub-critical everyday locations where most people spend most of their day, and especially where vulnerable people spend their day. The most vulnerable members of society are not productive in the conventional economic sense. This means they tend to have far less resources spent on them. The result? Commercial offices of million-dollar corporations tend to get the best HVAC has to offer. Public schools do not.
In a 2024 edition of The Conversation, Geoff Hanmer argued that the economic case for equipping every Australian school with air conditioning, at a cost of maybe 2 billion dollars, was a no-brainer. A professor of Architecture, director of OzSAGE, managing director of his own architectural consultancy, Geoff knows what he’s talking about. I have heard Geoff speak a few times. He asks why it is we accept a second-rate solution for schools, and it’s hard to answer.
This second-rate solution tends to be relying on staff or students to open windows. As any engineer knows any engineered system is only as good as its human interface. In other words, the system generally works fine until you introduce people. The promise of automation is simple and seductive – reducing the reliance, as much as possible, of the system on people will make it more efficient and allow us to get on with other, more productive things. And, anyway, people are terrible at knowing what they want, or trading off their needs and wants with their neighbours. Let the machine do it.
Yet we’ve all experienced the HVAC system that doesn’t work, or doesn’t get fixed for months, or may or may not work – nobody really knows. But don’t worry! AI is coming to the rescue! By studying us, AI will be able to anticipate our every foible, bias and quirk, and serve us air perfectly tuned and tailored to our appetite.
Perhaps.
But when? When can I expect such a system to be installed in my son’s classroom?
When COVID-19 struck we didn’t need a solution to be rolled-out over the next 30 years. We needed it now! We didn’t have one that could be deployed in days. And we had little to no idea what the state of ventilation was, not just across our schools, but all across our entire national portfolio of buildings. So, using the precautionary principle, we locked them all down. When schools were re-opened in the endemic, but not yet fully-vaccinated world our government came under pressure to do something to protect staff and students. After much debate over their efficacy, portable air filtration units were bought for 1 in 3 classrooms across NZ. Minimal instruction was provided with them. In other words, they were intended to be used, as much as possible, in the classic manner of automation – switch on and forget about it.
The result? Well, strictly speaking we don’t know. The usage and impact of the filtration units was not tracked (even though it could have been). But I visit a lot of schools and know a lot of teachers. I always ask about their experience of using their filtration unit. Some of the most common responses are along the lines of “I don’t know if we had one”, “It never got switched on”, “It was too noisy so we removed it”, “ It was put in a cupboard somewhere and never came out”.
However, there was a second parallel initiative. Thousands of handheld CO2 monitors were also distributed to NZ schools. Again, neither the raw data, usage or impact of these devices was monitored – a huge missed opportunity in my view. But, this time, the intent was the opposite of automation. It was to engage.
The point was that although relying entirely on windows for ventilation is certainly inefficient, it is potentially very effective. It is also very inexpensive and can be implemented immediately. The key was to not use monitoring in a passive manner – as you would as part of an automation system – but actively. In other words, to guide and evaluate action. What these monitors achieved was to help some teachers to discover their own agency and build up some rudimentary air quality expertise. Not from reading, or attending a training course, but through real-world trial-and-error experience.
The great irony, of course, is that the key to success was to use good educational practice. Happily, this particular penny dropped for a few teachers. These are the ones who turned this additional work burden into a natural part of the main purpose of their job – teaching and learning. Once monitoring was adopted as a learning activity for children, schools often saw rapid and major improvements in classroom air quality.
Teachers understand that we learn by doing. We learn from rapid feedback. We learn from taking rewarding excursions out of our comfort zone. We consolidate our learning by sharing it with others. Teachers know that we all yearn for a sense of competence, and ideally mastery. We don’t always, necessarily, want the machine to do everything for us.
An educated and engaged populace has a better understanding of what they want, and how trade-offs may be needed to get it. They are much better at improvising when the automated system goes down, or is just off doing its own thing. They are better at shifting to a new normal when the world changes.
So, from around 2023 we started re-thinking how to infuse classroom air quality management with good educational practice – designing the system from the teacher’s perspective. We’ve recently finished a small-scale trial of a “smart” air quality management system in a primary school in New Zealand (see our article and video here). Despite having monitors, remote-control filtration units and multiple feedback loops it most definitely is not about automation. Rather it is a deliberate attempt to maximise human involvement in the system – not by telling them what to do, but by providing a platform to help them learn the impacts of their actions. It’s also designed as a learning educational experience that gradually empowers children to handle more and more complex multi-objective conditions. After a couple of months we found our participants – 9 to 11 year olds – experts at how many windows should be opened how much, and where the filtration unit should be placed running at what speed, to balance CO2, PM2.5, temperature, noise and heat loss. Perhaps a super-smart AI-driven system could deliver the same environmental outcomes? But could it also deliver a cohort of children who hosted a dynamic event at school to educate other classes and their families, and who will probably carry their knowledge and drive with them for life?
I can’t claim that an education-oriented approach is superior to an automated one. And what automation can deliver is clearly going through something of a revolution at the moment. The point is probably that it is not a case of either-or. We almost certainly need both. But if an education-centred approach has a home, it must clearly be within educational establishments. But schools are not the only places where we learn, but merely the place where we ideally start our lifelong learning journey.
Oh, and automation (AI) was not used in the preparation of this blog.
