As African militaries increasingly explore AI to boost their capabilities, experts say that it’s crucial to remember the most important element in any AI system is the human element.
“The human being needs always to stay at the center of things — in the loop,” Col. Abdel-Aziz Fall, commandant of the Signal Corps Training School for the Armed Forces of Senegal, said during a recent webinar hosted by the Africa Center for Strategic Studies.
Fall and counterterrorism expert Niccola Milnes of NVM Consulting shared their expertise during the webinar “Artificial Intelligence and Counterterrorism.”
“We need to understand that AI is a tool,” Fall said. “It helps us with decision-taking, but it’s not the last decision that’s taken.”
At the heart of the discussion of AI and counterterrorism is the idea that AI is a powerful tool for analyzing massive amounts of data quickly, but it relies on human beings to have both the skills and the knowledge to write algorithms and to act upon the results of AI analysis.
“AI will never replace people,” Fall said. “People who know how to apply these skills will have an advantage against adversaries.”
AI’s current strength lies in analyzing imagery and recognizing patterns, Milnes said. Those capabilities can also help human users predict actions terror groups might take in the future. Much of that surveillance data comes from drones launched to detect terror groups as they move about the landscape.
“We can find activities that we never would have identified without those drones,” Fall said. “There is value in the rapid transformation of data into useful intelligence.”
Even as governments deploy AI against terror groups, those same groups are likewise using AI for their own purposes. During a recent attack on a Malian military outpost at Sévaré, for example, JNIM appears to have used hovering drones and AI to quickly recalibrate its targeting, according to analysts. The attack killed Malian soldiers and mercenaries with Russia’s Africa Corps.
JNIM also appears to be using AI to compose propaganda it posts to social media.
Drones, satellites and other technology make it possible for governments to collect vast amounts of data of all kinds. The first step in deploying AI for data analysis is to understand the data the government has and the capability that currently exists for making sense of it, Milnes said.
As militaries gather data, it’s also vital that they retain control over it as much as possible. That means developing their own digital systems for storing it and analyzing it whenever they can. Using foreign software or another company’s proprietary technology could create questions about who owns that data, Milnes said.
“Governments need to prioritize sovereignty over their data,” she said.
Homegrown capabilities will allow African nations to respond quickly to incidents in the field rather than having to wait for an answer from a foreign tech company, she added.
“If a problem comes up in the field, it needs to be solved immediately,” Milnes said. “There’s no time to wait for approval from higher-ups.”
Building local AI capabilities is a key part of keeping data safe. It can also make procurement simpler while engaging with homegrown technology companies, Milnes said.
That’s why the first challenge of any AI project is ensuring that people have the skills to manage it properly, Fall said.
“Without the skills, you can have the best platform, but it doesn’t serve you,” Fall said. “We can’t just buy the technology that we don’t understand.”