New DND strategy warns Canadian military's approach to artificial intelligence 'fragmented'
Strategy, which recommends internal AI department centre, gets mixed reviews
Canada's defence department and military are approaching the implementation and use of artificial intelligence in a "fragmented," unco-ordinated manner, says a newly prepared federal strategy.
The long-awaited overview of how to deal with rapidly evolving machine-learning technology is being met with mixed reviews by some high-tech and civil society experts who had been expecting more clarity on where the country stands and how far it's prepared to go.
The strategy is candid, saying neither the Department of National Defence (DND) nor the Canadian Armed Forces is "positioned to adopt and take advantage of AI."
The report described initiatives as "fragmented, with each command and environment addressing AI independently" and expertise scattered throughout the institution.
"No roadmap exists to move the organization toward leveraging AI effectively to ensure that investments are co-ordinated and appropriately governed, or to develop the capabilities, attitude and skills to implement AI effectively, safely and responsibly," the report said.
The document, which was recently released with no public fanfare, recommends following the lead of allies by establishing an internal defence department centre that "will act as a hub of AI expertise and an accelerator for experimentation, testing, evaluation and fielding AI."
The Royal Canadian Navy has been at the forefront of experimenting with the technology, asking Defence Research and Development Canada Centre more than six years ago to set up an AI system that could predict machinery failures on warships, the strategy noted.
"The [AI] system's performance was not perfect, with false positives also produced, but the initial results were promising enough to warrant further testing," the document said.
Allies such as the United States, Britain, Ukraine and Israel, to name a few, as well as adversaries such as Russia and China, have been incorporating the technology into various elements of military operations and planning at an extraordinary pace.
For example, in late 2021, the U.K.'s Royal Navy approached Microsoft and Amazon looking for a more effective way to co-ordinate military operations, according to several trade publications.
Within three months, the tech giants, a major British arms manufacturer and Ministry of Defence officials received a demonstration of a proposed AI solution, known as Project Stormcloud, which automates tasks and flags data for intelligence analysis.
Release of AI strategy a start, expert says
Branka Marijan, an expert with peace research institute Project Ploughshares, said she had expected the Canadian strategy to be more forward-leaning, given what allies have already learned and the real-work battlefield experience of countries such as Ukraine.
The Liberal government's recent defence strategy mentioned AI at least nine times, she said, but there appears to be a disconnect between what Ottawa says it wants to do and what it is doing.
"This strategy is more of an overview of ... things that we are thinking about, considering, than about possible applications of military AI," Marijan said. "There is a bit of a lack of vision in terms of what they want to do."
On the flip side, she said, it's good to see Canada being among the first countries to release an AI strategy for the military. But it needs to be more comprehensive beyond acknowledging that human beings need to be the decision-makers when it comes to the battlefield and support applications of the technology, Marijan said.
Canada, with its sophisticated high-tech industry and research, can play an important role in international discussions about which AI systems are acceptable and where the red lines should be drawn.
"I think what we really need to address at the national level is exactly how we see Canada moving forward with the application of these technologies," Marijan said.
"I think there are important decisions that are going to be made about which systems we decide to deploy and which ones we do not.
"In the democratic society, we really need to know how we're using AI to aid decision-making, especially in defence applications. This is a context where we need a lot more transparency and oversight, which we don't usually get."
There are also practical considerations and questions about whether the country has the digital talent for the military to draw upon, Marijan said.
Daniel Araya, a senior fellow with the Centre for International Governance Innovation and a noted expert in AI, said he has no worries about the talent pool. Rather, he said, success will require the military to rethink the kind of people it recruits.
"It's not that we don't need infantry, or we don't need people that can fight with weapons," Araya said, "but the fact is, we're shifting toward automation, augmentation. We need a different calibre of soldiers. And I think there's probably room for a complementary system between the fighter on the ground and the augmented software developer that provides a platform that makes the whole thing work."
The DND will have to be very careful and deliberate in its choice of which private-sector companies to partner with, Araya said, and those relationships need to include Canadian companies — not just the big U.S. tech manufacturers — because "this is about ethics of use of force, ethics of the application of AI and machine learning."
The approach to integrating the technology into the military also cannot be treated like a standard defence procurement, he said, and "AI is not something that can be bolted on" to existing procedures.
In addition, Araya said, the DND has to realize that it's on the leading edge of a revolution that will touch many aspects of life outside of the military, so in many respects it will be a role model for governance and innovation.
"It's like electricity," Araya said. "It'll be deployed in the military, it'll be deployed in health care, in education, and transportation and communication. It'll be everywhere. And so by building AI capacity that serves the military, we also build AI capacity. It serves all our other industries."
ABOUT THE AUTHOR
