DLA CIO discusses logistics, where AI potential meets reality

The Defense Logistics Agency's chief information officer cautioned that successfully integrating artificial intelligence requires more than just finding a use case and deploying a new tool. It demands a significant investment in the workforce and a shift in organizational culture.

"Not every problem is an AI problem. We need to answer where AI fits in compared to business processes or versus automation," said DLA CIO Adarryl Roberts during an artificial intelligence innovation panel at the National Defense Industrial Association Annual Logistics Forum in Tampa, Florida, Feb. 17.

A certain level of data and technology acumen and data quality is a prerequisite for the development of a data savvy workforce to effectively leverage AI, he said.

"Don't wait for perfect," Roberts said. "We are already making decisions based on imperfect data, so don't wait for perfect data to make a perfect AI."

He advocated for viewing AI as a tool in a constant state of self-improvement, similar to how current AI models are used to create even better future versions.

This ongoing development requires a new level of trust between human employees and their digital counterparts, Roberts said. While automated bots can perform the equivalent work of more than a dozen employees, AI systems are not yet completely trustworthy. To bridge this gap, he said that as AI models are trained to be digital employees, human employees must be trained to appropriately trust and collaborate with them. The goal is to see AI as a decision enabler, not a decision maker, capable of augmenting human abilities in ways previously not possible.

Roberts highlighted the agency's use of AI in predictive modelling as a strength. He also noted the increasing importance of data standardization as a key component of protection in the competitive AI battleground. He specifically mentioned the challenge of developing answers to a disconnected environment where adversaries can target connectivity as a means of disabling AI-driven systems and tools.

Bringing it back to DLA's core mission, he said focusing on the end users is critical.

"Stakeholder alignment pales in comparison to the needs of the warfighter getting the material they need," he said.