The Nelson Artificial Intelligence Institute (NAI)
We live in the information age. All around us, sensors are sensing, networks are pinging and, increasingly, machines are learning. We are now generating so much data that according to the World Economic Forum it’s estimated humans will create 463 exabytes - or 212,765,957 DVDs - of data each day by 2025.
Humans are incapable of processing this much data. And that’s where artificial intelligence (AI) comes in. It provides a way to sort through all the information and find trends, patterns or insights and then use them to help solve problems.
“The idea,” says Dr Julian Maclaren, one of the co-founders of the Nelson Artificial Intelligence Institute (NAI), “is that we want to employ local people, solve local problems but find solutions that can be scaled globally.”
The NAI received a $3.4 million loan from the Provincial Growth Fund in 2019 to get going and initially focused on using AI to solve problems in the aquaculture sector, but it has expanded its remit since then and now focuses on three main areas relating to the environment: climate change, healthy oceans, and conservation and biodiversity.
Targets were set at the time that required a certain number of jobs to be created and, with around 25 staff between NAI and NAI-spinoff CarbonCrop, Maclaren says it has easily met those.
NAI is partly what it calls a ‘venture studio’ and uses a stage gate system to filter ideas.
“We put a lot of ideas into the top of the funnel, they drip through the stages and they get shelved if they don’t make it through or become separate entities if they do.”
So far CarbonCrop, which makes claiming carbon credits easier for landowners and aims to “pull a billion tons of CO2 out of the atmosphere and lock it up in trees,” is the only idea to make it through that process, so it is now a separate entity and has received private investment to take it further.
Nick Butcher was the ‘entrepreneur in residence’ who was committed to that idea and he has now become the co-founder and CTO of CarbonCrop.
“We are looking for more people willing to drive these environmental AI projects,” Maclaren says. “It’s a bit different to an incubator. Incubators would take less equity but they provide less. We provide everything, including a salary. And maybe even the idea, and maybe some initial proof of concept. We have an engineering team who can help prototype and we’re well-connected to funding.”
That funding is a crucial part of the process and he says New Zealand is now well-and-truly on the radar for those in the global tech industry who are looking for companies to invest in. That has come as a result of a few things: the success of some of our high-profile tech companies and a range of promising start-ups; the tech elite’s enduring fascination with New Zealand; and a whole heap of positive global attention after our Covid response.
In 2012, Maclaren moved from Germany to California, where he worked on projects at Stanford University, GE Healthcare and Google but he and his wife decided to move back to New Zealand (alongside setting up the AI Institute, he is still working remotely in the US). While he grew up and studied in Christchurch, they picked Nelson.
“It’s a great place for kids. The weather’s good and is about as similar to California as it gets in New Zealand. And there’s great access to national parks,” he says.
Nelson may be close to California when it comes to weather, but there’s a fair way to go when it comes to tech jobs, and that’s what NAI is hoping to change.
“There’s one guy who is originally from Nelson. He got a degree in engineering, went overseas, studied machine learning in Berlin and now he’s happy to be back here.”
“It would be great to grow more talent in Nelson eventually, but realistically the people we’re hiring, they have quite specialised skills and they need a lot of experience. To me, it’s more about bringing back New Zealanders who have gained that experience, or attracting people here from overseas who can help.”
One of NAI’s directors, Alexey Rostapshov, arrived in New Zealand a couple of weeks before lockdown in 2020. He works for John Deere Labs in California and was part of the Edmund Hillary Fellowship programme.
“He’s still here. And he’s been incredibly helpful. There are other people like him who are thinking ‘maybe we should relocate to New Zealand.’ I think he’s been pleasantly surprised. He certainly hasn’t been rushing to get back.”
One issue holding some people back is the partner factor, which is often pointed to as an impediment for a region like Nelson Tasman.
“The person you hire may also need to consider a job for their partner. We’re quite specialised, so it’s not likely we could offer them a job so it’s really important to us to see what else is going on in Nelson. Could we put together an email list for that particular problem where we can direct CVs?”
When Maclaren was asked to work at Stanford, for example, one of the first things they asked was, ‘what does your partner do?’
“They wanted to help. It's expensive to live in Silicon Valley and it’s what they do because they’re part of a big community. You have to do it if you want people to relocate.”
Wages are also an issue. There’s huge demand for AI experts all around the world, so if we want to attract them to this part of the world he says we need to pay them market rates. Living expenses are lower than big cities and there are many advantages to living here, so wages can be slightly lower, but he feels the answer is to think global from the start.
“We have to sell to the world from here. It’s easy to say, but it’s hard to execute.”
Fortunately, it’s never been easier to do it because, last year, video meetings, online communication and dispersed employees became accepted as the norm.
“What we’re building is potentially usable globally with a few software changes where we train AI models to work in other locations. We don’t need to physically ship anything.”
Tech companies like Google, Apple and Facebook like to talk about ecosystems. In some cases, the academic institutions around those tech hubs do the research, and that then flows out into the private sector. NAI works closely with the Nelson Marlborough Institute of Technology and Brian Russell, one of the founders of NAI and a senior biomedical engineer with NASA Ames Research Centre, worked with NMIT to develop course materials. NAI is also sponsoring AI projects being done by a number of students at the University of Canterbury.
“Our first collaborator was the Cawthron Institute. This has been great because they have domain expertise and we have the AI expertise. Often the problems come from a specific domain and then we can tailor a solution to that.”
So far NAI has used technology to help detect whales and other sea life, automatically identify microscopic organisms or materials, and spot possums, cats and other predators on video footage.
Local iwi and organisations like Wakatū are increasingly bringing attention to the importance of long-term thinking and he hopes the Te Tauihu intergenerational strategy and principles like kaitiakitanga / guardianship and reverence for the land are starting to flow through to the wider business community.
He believes there is now more willingness from companies to find ways to reduce their impact on the planet, both because it’s the right thing to do and because it’s the most efficient thing to do.
“The timing is definitely right in terms of carbon. There’s a price on it, which helps, but there’s also finally widespread acceptance that climate change is a major problem.”
But AI is not some magic box that will solve all our environmental problems. It’s a data automation tool that can help solve particular problems, he says. But you can’t manage what you don’t measure. And Maclaren and the team are doing everything they can to help measure things from Nelson and find solutions for the world.