The Air Force’s Five Principles to Advance Artificial Intelligence

The 2019 NDAA bill provides for the establishment of the National Security Commission on Artificial Intelligence, which will explore how the military and Department of Defense use and develop AI technology. (Sean M. Worrell/Air Force)

The 2019 NDAA bill provides for the establishment of the National Security Commission on Artificial Intelligence, which will explore how the military and Department of Defense use and develop AI technology. (Sean M. Worrell/Air Force)

July 25, 2019 | Source: C4isrnet, Kelsey Reichmann, 26 June 2019

"The Air Force has been on an almost 3-year journey to integrate artificial intelligence into operations, and that effort will soon be more apparent as the Service plans to declassify its artificial intelligence strategy," Capt. Michael Kanaan, the service’s co-chair for artificial intelligence, said June 26 at the AI World Government Conference in Washington, DC.

“We had to find a way to get us to a place where we could talk about AI in a pragmatic, principled, meaningful way,” said Kanaan.

During his speech, Kanaan laid out five principles that have guided the Air Force with artificial intelligence in the meantime. They are:

1. Technological barriers will be a significant hurdle.

Kanaan said the service has made it a point to limit technological obstacles. However, one problem contractors may face is higher priced products geared toward security-driven government programs vs. the same, less expensive commercial programs. "A new attitude toward commercial off-the-shelf technology within the service can help," he said.

“Too often, working with our agencies, they have to take risks in the framework of time, people, and bespoke unique solutions to deploy on your systems,” Kanaan said. "However, this does not have to be the case. Accept commercial standards because unclassified does not mean un-secured."

2. Data needs to be treated like a strategic asset.

“We used to ask the question, if a tree falls in the forest does it make a sound. Well, in the 21st century, the real question to ask is was something there to measure it,” he said. He explained this involves looking at when and how to digitize workflows.

3. The Air Force must be able to democratize access to AI.

“This is an opportunity now to say machine learning as our end state, if done right, should be readable to everyone else,” Kanaan said. "This will involve balancing support and operations and taking into consideration the reality that the demographics of the traditional workforce are going to shift," he explained.

Communities: