Experts from government and academia said this week that successful adoption of artificial intelligence (AI) technologies depends on a host of factors including strong leadership on the technology front, along with a good understanding of the contents of data sets that are used to train AI applications.
Technology leaders, said Dr. Brian R. Spisak, consultant and research associate at the National Preparedness Leadership Initiative at Harvard University, need to adopt a “lead first, tech last” approach to AI development.
During a Jan. 9 panel discussion organized by ATARC, Spisak argued that tech leaders must first understand their organizations’ AI needs before jumping on the latest technology trends.
“Leaders need to dive really deep into factors like workflows, employees’ behaviors, and things like data quality, and I can’t stress enough the importance of data quality,” he said.
“Once you have an understanding of your people and those workflows and processes, then you can start moving into that strategic planning area and that space where you’re trying to think about how to optimize processes,” added Spisak.
Dr. Frank Tortorello, a cybersecurity analyst in the office of the U.S. Senate Sergeant at Arms, talked about how AI applications can streamline some of the “boring” tasks that take up valuable workforce time, and free up time for people to take on more important work in a safe manner.
“For me, the logic points to the idea that fully exploiting the opportunities offered by AI and maximizing risk reduction means putting time and money and leadership … into understanding and developing the abilities of people to use the tool safely,” stated Tortorello.
Dannie Lyvers, Insider Threat Program senior official at Georgia Institute of Technology, talked about the importance of understanding the contents of data sets that inform AI tools before they are deployed in order to protect data that makes up the “crown jewels” of organizations.
“We’re taking a look at the protection levels, on how AI can come in or the machine learning aspects supplement what the human aspect is getting and collecting data … because you don’t know, what you don’t know until you actually go out there and you find it and collect it,” stated Lyvers.
