Hollywood bound: Aussie tech start-up Fire Foresight uses AI to get a jump on bushfires

Fire Foresight has attracted the attention of Los Angeles mayor Karen Bass after it built a new platform to better detect bushfires, thanks to artificial intelligence.

Fire Foresight's system installed at Mt Horror, Tasmania. It uses an AI-powered platform that can detect faint smoke signals in real time, often before they are visible to the naked eye.

An Australian start-up has attracted the attention of Los Angeles mayor Karen Bass after it built a new platform to detect bushfires earlier, thanks to artificial intelligence.

Fire Foresight is the latest tech company to erupt from Tasmania’s burgeoning start-up scene. It has built an AI platform that can detect faint smoke signals in real time, often before they are visible to the naked eye.

And it’s already in action. During last year’s bushfire season in Tasmania, the Fire Foresight system detected about 500 blazes – sometimes up to 15 minutes ahead of other alerts.

Fire Foresight founder and executive director Rob Vernon.

Founder and executive director Rob Vernon said in 90 of those fires, his platform was the first or provided the only escalation to a fire management agency.

Now California – the tech capital of the US – is looking at Fire Foresight’s technology, following its wildfires earlier this year, which caused Allstate Insurance estimated cost it $US1.1bn ($1.69bn) worth of devastation and other estimates have placed the damage bill at more than $250bn overall.

Mr Vernon visited Ms Bass in Los Angeles after attending the South by Southwest festival in the US this year.

“I was at the mayor of LA’s place talking to her about what we do,” Mr Vernon said.

“They do have a large network of cameras – something in the order of 1300 cameras – up the length of California. But they (the mayor’s office) have been working with their partners on trying to bring AI to that game at scale.”

Mr Vernon is in the early stages of planning a $2m capital raise to put more “boots on the ground” and “compress the sales cycle”.

“The technology is under control, we’ve been bootstrapping to date. But we will need to raise in order to do that at scale in Australia and where we want to go.

“I’ve been having a lot of conversations with investors … while we gain more revenue and traction. We’ve got lots of opportunities.

“The average deal takes about 12 months to get across the line. We’ve been doing this for a bit over 12 months, so that pipeline is about to start converting.”

One of Fire Foresight's cameras in action in Tasmania.

Tasmania is emerging as one of Australia’s hubs for tech talent. Canva co-founder Cameron Adams moved to the state, as have hoards of Atlassian employees, while ASX-listed ReadyTech launched a centre of excellence last year to create more training and career pathways in the sector.

Mr Vernon, originally a software developer, has a 20-year career – mainly in telecommunications, including serving as chief executive of TasmaNet.

Fire Foresight has trained its own AI model to detect blazes.

“We used to use a model that’s been trained up on just globally available smoke imagery. So it’s a computer vision model. The challenge with that is it doesn’t necessarily look like Australia all the time, let alone Tasmania. So we now train our own model.

“We’ve got enough source imagery of what is fire and what isn’t a fire from our work down here. So we’re just continuing to build up our own models based on that.

“Validating and verifying that by incorporating other sources of data is a critically important part of the AI platform as well, so that we’re getting humans involved in the decision making.”

Fire Foresight founder Rob Vernon said his system could have given ‘minutes worth of heads-up to humans’ during the California fires earlier this year. Picture: Getty Images

Mr Vernon said his system could have given “minutes worth of heads-up to humans” during the California fires earlier this year. If the conditions were not as extreme it could have provided more notice to authorities.

“It’s all about just super early detection from a human safety point of view. Now, had the conditions been a little milder or little less (severe), then again, the same thing plays out.

“But the earlier you can get on top of a fire, the greater your chances of actually being successful putting it out.”