Bushfire Social Intelligence – 2023 Michael Bryce Patron’s Award



In the summer of 2019-2020, an estimated 16 million hectares of land burned across a drought-stricken Australia. Aptly coined the Black Summer, hundreds of blazes ravaged ecosystems, communities and infrastructure for almost twelve months, taking the lives of over 30 people, killing or displacing up to 3 billion animals and causing around $200 billion in damages.

In this catastrophic summer, more individuals than ever before forgoed the traditional “000” route to instead take to social media to share bushfire updates, stay connected and cry for help. It saw emergency services needing to quickly adapt to a new online norm, sifting through status updates and photo uploads to piece together bushfire trajectories and strategise their next moves. 

The people, processes and existing technologies of Australia’s fire services were pushed well beyond their limits as smoke suffocated the country. So, in the wake of a scorched continent and a radical shift in the way individuals report bushfires, Abby Phillips – Senior Product Designer at Kablamo – and her team sought to prevent such a catastrophic event from ever happening again.

The result is Bushfire Social Intelligence, a world-first emergency services workflow providing highly relevant public information to firefighters. The machine-learning (ML) and artificial intelligence (AI)-optimised innovation assists in early fire detection, fire behaviour analysis and response time enhancement. Bushfire Social Intelligence has also been named the 2023 Michael Bryce Patron’s Award recipient. 

Good Design Australia caught up with Abby Phillips and Dee Behan – Managing Director and Design Principal at Kablamo Design Studio – to dive deep into the transformation project, its potential and the importance of human-centred design. 

Bushfire Social Intelligence – 2023 Michael Bryce Patron’s Award

Good Design Australia: Australia is a continent known for the bushfire, with many fire-led catastrophes dotting its history. How would the Bushfire Social Intelligence platform come into play in the time of a fire?

Abby Phillips: Bushfire Social Intelligence includes a suite of features integrated within Athena – a platform that we’ve been developing with the New South Wales Rural Fire Service for nearly two years now. Using a ML model trained on keywords and hashtags that are relevant to bushfires or emergency services, it collects and visualises vital information. It also uses imagery recognition, looking at imagery on social media that has fire, smoke and other landmarks.

All relevant information can be plotted on a map within Athena which can show its relationship to known incidents, whether it’s near a known fire or a hotspot, or even infer that there is an ignition we didn’t know about. Emergency services all have teams that sit at computers and trawl through Facebook, Twitter and Instagram looking for those posts, so Bushfire Social Intelligence speeds up the process to detect or slow down fires before they get out of control.

GDA: It seems a lot of raw data, keywords and photos are collected in the social media scanning process. How can the platform tell what is relevant and what is not? 

AP: That’s the machine learning part, where there are always challenges to overcome. A lot of computer vision models that already existed were based on the US context, so we had to spend time training the AI model in the Australian context with the Australian environment, and we still have more work to do in this space. With the keywords and hashtags, something that became obvious in the design process was there are a lot of keywords that trend due to different societal issues at a point in time that do not contribute to bushfire management. This means it’s not as simple as pulling a bunch of data from Twitter, for example, that was trending at the time and putting it into a machine learning model. 

So it was a lot of trial and error to optimise that, a lot of workshops, a lot of sitting down with users and people that had been managing social media data throughout recent bushfire seasons and understanding to them, what they considered to be data that is used as intelligence and not just general information generated from the community.

Bushfire Social Intelligence – 2023 Michael Bryce Patron’s Award

GDA: Would you describe this as more of a human-led approach to a ML and AI-optimised solution? Why is this important?

AP: Yes, for sure. I think if you’re starting with the technology and the solution first, then you’re not considering the problem you’re actually solving for. For machine learning in particular, I believe it only works if you’re starting with a problem in mind. That human-centred design mindset and the design-thinking approach really helps you uncover the problems you’re going to solve, who they are for and whether or not we can solve them with these people. Not for them or at them, but bringing them on the design journey.

For the Bushfire Social Intelligence project, the problem we were primarily looking to solve is the massive amount of time emergency service personnel were taking to sift through all of this information, and then the time it takes the information to get to the people that need it to make those really critical decisions. Once we had defined that, we could understand that machine learning was one of the right solutions to apply to this problem. 

GDA: Talking scalability, can the Bushfire Social Intelligence technology be applied to other natural disasters for example? How can it keep up with the speed of social media?

AP: It can be applied to any natural disaster, as long as we’re working alongside people that are experts in those emergencies and understand what they’re looking for. After all, we’re not firefighters, right? We’re technology and design experts. So, we have to be designing with those people that have got the knowledge. We have a great baseline for what machine learning and object recognition models can do – it just needs the relevant data.

Talking about social media, we’ve actually designed the user interface and the technical architecture so that it can integrate with future social media or crowdsourcing platforms. This sees repeatable design patterns and content containers flex to allow data from any social media platform be used. It means we don’t need to spend time redesigning an interface to visualise information from new sources, and can focus instead on other high impact initiatives.

GDA: Zooming out a bit, how would you describe design’s general potential to face or challenge big societal or environmental issues? In what ways can it ignite positive change within communities?

AP: I think that, as designers, we know how to ask the right questions. We’re really good at being strategic thinkers and thinking of the problem holistically. We can partner with the users, and the technology experts – we’re really the people that are able to join the dots between the whole picture. Design can create an interconnected, innovative environment that allows innovative technologies to be applied to the problems we’re solving.

Dee Behan: Just to add to that; with any sort of societal change or environmental change, there’s always those systemic barriers or legacy systems that we have to overcome. Without a design-thinking mindset, it’s very difficult to just inject a new way of working or expect a new tech to actually be adopted.

As Abby was saying, it’s going right back to all that interconnectedness. Designers can really map together and understand how a positive change can happen. It works. Just looking at the [Bushfire Social Intelligence project], the work that the team is doing is ultimately saving lives, saving our environment, it’s saving animals. It’s actually the most crucial work. I think the vision at its core is that “we will be saving lives”. That is at the forefront of everybody’s mind.


More News