Agriculture warned of artificial intelligence challenges

Glacier FarmMedia – An expert told a conference of agricultural leaders that Canada must deal with a crisis in public confidence in artificial intelligence, likening it to the monster under the bed.

Surveys have shown many Canadians don’t understand AI, and how it can provide a leg up for industries such as agriculture, said John Weigelt, national technology officer for Microsoft Canada.

“And so, in a sense, it’s the monster under the bed. We don’t know what’s there, we don’t know what it is, but we’re scared of it.”

He spoke at the recent GrowCanada conference in Calgary as part of the Agriculture 4.0 panel, attended by several hundred people. The annual conference, which alternates each year between Calgary and Ottawa, included leaders ranging from farm commodity organizations to life science, seed and data companies, as well as the banking and finance sectors.

When panel moderator Shaun Haney asked about who in the room was excited about AI, many people in the audience raised their hands. However, a fair number also raised their hands when he asked if they were “bloody terrified.”

Weigelt described AI as the ability to reason over a set of data.

“I think that’s the simplest way to put it, and so that data might be an image, that data might be audio, that data might be just statistics that you have. And so, I lean in a little bit on this idea of data really changing the nature of how agriculture does its business.”

As someone who is accountable for responsible AI within Microsoft Canada, Weigelt said Canadians lag in the adoption of AI tools, “so, we need to scratch the surface. We need to understand that AI depends on data, there’s math involved with it, and then there are tools like responsible AI to make sure that, hey, there’s no bias, that you can trust it, it’s reliable, secure, that it’s inclusive, transparent and all those things so that you have the confidence that these things are deciding correctly — that they’re doing what you expect them to do.”

He pointed to the recent brief removal and reinstatement of Sam Altman as chief executive officer of OpenAI. As the leader of one of the most powerful players in AI, Altman was at the centre of a dispute by the company’s board of directors about the power and future of the technology.

A new era of AI began after ChatGPT was launched last year by OpenAI, reaching about 100 million subscribers only two months later, “the fastest adoption of any tool set ever before,” said Weigelt. People can now use computing resources as easily as they might talk to their sister or neighbour, he said.

“That’s the allure of this, that you can have that interaction and use computing resources that quickly and easily. Now, we can’t get them to do everything, just like you can’t get your neighbour to do everything, so we have to find those spots to be able to do that.”

Weigelt has worked with a wide range of Canadian industries, including agriculture, to try to determine where the latest cutting-edge technologies are heading. They include advances such as synthetic biology to develop organic computers, “so think of using DNA to do computations,” he said.

“(We’re) talking about quantum computing, and how quantum computing can address things like climate change, address things like cracking the nitrogen cycle so that we can feed more people. And looking at older technologies like cloud computing and helping Canadian organizations move forward with some of these new paradigms.”

Cloud computing has allowed companies to do things over a weekend that used to take months, allowing them to adopt a mindset of innovation that operates on a much faster timescale, said Weigelt. However, although Oct. 27 marked the 15th anniversary of cloud computing, adoption remains low in Canada, he said.

“And so now, when we see these things coming together — we see the drones. We see the Internet of Things, we see that coming into the supply chain, so we have end-to-end visibility from production through to consumption. We see that in the AI becoming involved as well. These things are all accelerating the change.”

Such technologies could allow the agriculture industry to address problems ranging from climate change and drought to labour shortages by promoting innovation through productivity, said Weigelt.

“And if we as Canadians aren’t adopting these tools and using these tools, then we are not going to be able to take advantage of that change. We’re not going to be able to seize that prosperity and see the benefits that were mentioned earlier around addressing the crises that face us.”

Weigelt was Microsoft Canada’s representative in the development of the federal Global Innovation Superclusters. The billion-dollar initiative was launched in 2017 to bring together companies, academic institutions and not-for-profit organizations to boost growth and innovation in five key sectors.

They include the Prairies-based Protein Industries Cluster, which aims to increase the value of key crops such as canola, wheat and pulses. It also strives to better serve markets for plant-based meat alternatives and new food products in North America, Asia and Europe, said a federal website.

However, Weigelt said he encountered obstacles when he started to have conversations about data for the cluster.

“Well, the first thing was they didn’t let me in the room. I wasn’t from around here, so that was kind of awkward for me.”

It also turned out there were what he called a “bunch of silos” preventing people from accessing data.

“Soybean people didn’t talk to wheat people didn’t talk to corn people, and yet it’s the same soil that they’re using.”

Although Innovation, Science and Economic Development Canada wanted to create projects that would float all boats for proteins-based innovation, it wasn’t able “to get that data so that you can reason over the data and make sense of it,” said Weigelt.

“And so, we really need to overcome some of those hurdles to be able to start sharing the data, making sense of the data, so that you can reuse that data across other places. There needs to be a governance model, and the governance models are coming into play so that the owner-operator can then own that data or monetize that data.”

– Doug Ferguson is a reporter for The Western Producer.