In 1955, General Electric invented GERT. GERT was a computer and at the time the fastest in the world. Today, your phone is 927 million times more powerful than GERT. In 1955, people held hoses and pointed them at fires. We are still doing this.
I’m not a firefighting expert, I’m a technologist. I have never been in the field face-to-face with a bushfire and neither has anyone on my team. We know technology and that’s where we can help.
After the Black Summer of 2019–20, my team and I committed ourselves to learning as much as we could about bushfire fighting in Australia: 33 deaths; 3,000 homes destroyed; 16 million hectares burned and 3 billion animals killed; $100 billion dollars in damage – almost 10% of Australia’s GDP. We felt strongly that the nation should be on a bushfire war footing, and we wanted to do our part.
First and foremost, what we found was incredible people, the firefighters, committed to being on-call day after day to protect lives and property. Australian firefighters are arguably the most experienced and the most tested in the world. They are faced with an incredible challenge, and what we also found was that despite Royal Commissions, multiple inquiries and endless political rhetoric, they are being let down and significantly under-resourced in critical and fixable ways. And because they are being let down, all of us are being let down and the clock is ticking to another cataclysm.
Technological advancements have simply not reached firefighting the way they should. This is a complex problem, and it will require collaboration by government, education, science communities and the private sector to solve. I want to outline the challenge facing us and then offer a sense of what a future state that has successfully integrated technology can look like.
The current technology gap
What’s the single biggest problem? It’s the lack of real time, data-led decision making. Many of our fire services still rely on manual and paper-based processes, and much of the decision making is still based on human intuition.
If you order dinner from UberEats what are the chances that the driver will go to the wrong address? Pretty much nil. Today, the consumer experience is largely digital and error free. At the enterprise level we know that we can now scale up huge data sets to meet demand.
But as revealed in the recent Royal Commission, some Australian fires are being caused by simple but catastrophic errors: people starting hazard reduction burns in the wrong areas guided by paper-based maps, or giving the wrong coordinates over radio channels. Here are two sobering quotes from the RFS investigation of Balmoral backburn gone wrong: ‘Commissioner Rogers said an internal investigation found the backburn quickly burned out of control because it was lit in the wrong location due to a mapping error.’ And a local RFS captain said this: ‘Given the forecast for the next day, which was 42 degrees, very strong westerly winds, it was just a terrible decision to start the fire 2.5 kilometres out from a village.’
These are problems that are solvable with technology that exists today.
At a typical on-the-ground command centre for an emergency response you’ll see whiteboards, lots of paper including paper maps but only one or two laptops. The glaring problem to those of us in tech is that the manual and the digital remain unconnected. What’s on the whiteboards and paper stays there or only later gets laboriously transferred into the digital.
Everyone involved in a firefighting effort is dealing with critical information that could be made available on digital interfaces and shared in real time, to make decisions much faster, with more precision and widely accessible.
The problem isn’t a lack of data to make those decisions but that the current technology platforms aren’t built to collect and share that data in a way that works for emergency response. Most of them have been built for fire scientists and fire analysts, often for academic purposes, without the intuitive user design and software development that would make them useful for people in the field.
They’re not fast or intuitive enough and, combined with manual processes, they’re struggling to cope with longer, more extreme fire seasons. We’ve already seen countless examples of the existing technology systems being overwhelmed because the complexity and scale of a fire was something that had never been encountered before.
Not only that, when systems are built without the user in mind, they are often left unused by the very people they should be serving in favour of a return to more established methods.
In our learnings over the past year, we have seen our firefighters pushed back to the paper-based, manual processes and decisions made on human intuition that might feel safer – but are, in fact, less effective and more prone to error.
The four data challenges
In Australia and around the world, large organisations and governments are building AI and machine learning engines to better utilise the deep pools of data they have. But what is it about bushfires that makes harnessing the data so much more complex?
Any technological solution has to overcome four challenges: variety, resolution, sparsity and latency.
First, variety: the almost limitless amount of available data that could be involved in making a single decision. The right tech platform must prioritise user experience to determine which data is most important, and when.
The second challenge is the resolution of the data: what level of detail can be captured, and how is it processed, searched and presented to the user? To map a large fire you could place a million different points around the fire at any point in time, with each moving every five or ten seconds. Actively making decisions with that resolution of details is unmanageable. Instead, the best solutions will capture data in the highest resolution, but search and surface it in a lower resolution.
The right technology must address sparsity. Bushfire data by nature is sparse. How do you interpret irregular data feeds and fill in the gaps? Individual data points can be intermittent – aerial scans or satellite images that come through every now and then, separate radio and 000 calls. New technology needs to effortlessly stitch together disparate data to create a clear picture of one event.
If the time lost to manual processing is one of the major faults of analogue firefighting, how can we minimise the time between data capture and computer processing of the technological solutions? While it might not rank highly for technological transformation in other arenas, latency and remote connectivity challenges are a priority for firefighting.
When my team was commissioned to develop an easily accessible and usable data platform for a state fire agency last year, we had to bring all of these elements together to successfully integrate our technology with their team. Each aspect of the digital transformation is important, but one is critical to mention here: the human behaviour factor.
Arguably the biggest challenge is this human element – changing behaviour is always much harder than upgrading the technology. There is no point investing in technology if you don’t invest in the organisational change capabilities to go around it. To successfully revolutionise firefighting, it begins with vision and leadership from the top. This is crucial to achieve buy-in and legitimacy for the next element: user-centred design. Tech systems should be built based on what people on the front line actually need, designed in collaboration with those very users. This means really listening to how they work – solutions pushed top down tend to fail.
Experienced change management requires communications, training and an ongoing commitment to making the changes stick, but above all fire agencies and organisations need to be willing to change their ways of working.
Different financial models or partnerships can unlock innovative solutions if we look beyond the traditional ROI models. For example, the recent partnership between Optus, the ACT Fire Service, Australian National University and Minderoo to install ground-sensing cameras to improve bushfire monitoring – if successful, the trial could provide a framework to be replicated in all states.
How the future could look
If Australia can overcome the data and transformation challenges that have confined us to manual-style firefighting, how far could the technological revolution take us?
I can’t help but make a comparison to the First World War – military technologies have advanced so much that it’s time we apply it to the way we’re fighting bushfires. Is it really necessary to be sending people to the front line of a fire en masse, the way we do now?
We have the capability and the technology for drones to play a part of that picture. While there are challenges around range, connectivity and the harsh, unpredictable weather conditions of bushfires, these can be overcome.
Elon Musk has also launched Starlink, with 1,300 low-orbit satellites already in play providing us with satellite-based high-speed broadband that can be accessed with an antenna. All of a sudden, we’re in a world where thousands of trucks can be connected. This connectivity can help us track not only assets like the trucks themselves but also people – the volunteers and firefighters – in real time.
But AI, drones, or new technologies won’t live up to their potential without data being harnessed and shared. No matter which of the many solutions are best for Australian fire services, integrated data platforms will be critical in unlocking the power of data, AI and robotics. The platform we built is a step toward a world where we have the information available to all of the different users – out in the field, in HQ, or out in public – being able to share and make decisions in real-time based on accurate data.
But to go forward into this future will take all of us working together. There’s no time to lose.
For more information, go to www.firetrail.io