AI maturation and adoption should have elements of a natural process to prevent system rejection and enhance cybersecurity.
Experts from the Department of Homeland Security (DHS) and the Office of the Under Secretary of Defense for Research and Engineering spoke today about that maturity process and how to maintain AI maturity over time.
When discussing a natural maturing of sophisticated analytics, Technical Director for the Homeland Security Advanced Research Projects Agency at DHS Steve Dennis said it must happen or otherwise “it may find space in the data center somewhere, it may actually be available, but what happens over time is that people stop using it, and you check in a year later, you find out that, ‘hey, this thing’s been turned off,’ or ‘it’s not cared for, it’s not working anymore.’” Dennis says it’s essentially just a great science fair project, if a natural maturing isn’t allowed to occur.
Not only does the AI model have to mature, but the workforce needs to mature as well. Dennis says there must be a focus on both cleansing AI data and the algorithms of the data, but that the people aspect is key. He offered that organizations need to embrace needing a workforce that has the capability to understand the underlying system and underlying models. “Just my own personal observation is that the government needs to think about how it’s going to organize around data,” he said.
Dennis mentioned that a top-down approach to this organization might not be the best way to approach the organizational structure around data. Using Uber and Netflix as examples, he explained that their organization’s models are “structured around improving the data” and improving data models.
