As the Federal government and Department of Defense look to find creative solutions to problems, creators will need to emphasize getting accurate nonbiased data and understanding the user experience (UX) of their product if they look to create something that is both useful and used, the Chief Data Officer for United States Special Operations Command (SOCOM) said.
Thomas Kenney – who holds a dual role as SOCOM CDO and Chief of Artificial Intelligence (AI) – emphasized the role both data and UX play in the content and product development process at the Adobe Experience Makers Government Forum on May 10.
Kenney said he’s driven by problem-solving and that, as he is looking to creatively solve problems using AI, the quality of the data is key to making solutions that work.
“People are saying, ‘Oh, we just want AI, AI, AI and we’ll all be amazing,” he said. “And the reality is … the data that you have, will have a direct correspondence to your effectiveness with which putting AI out there.”
Kenney referenced an industry project that looked to use AI to streamline the resume review process, and that after three years of work it ultimately had to be abandoned because it ultimately recreated prior implicit biases in hiring processes, as well as confirmation bias from the overseeing managers. He used this as an example of the need to carefully teach AI, as well as the need to start with more pedantic processes when implementing AI, such as annual routine paperwork.
“This is not the rocket science that we want to get to. But this is the busy work that keeps us from having the right budgets and the right opportunities to get after it,” he said.
“Imagine the amount of information you can gain by understanding your data and where it is and how you’re using it. And these are baseline concepts that if we can start with things like paperwork processes, people will become much more adept and much more comfortable delivering technology to automate what they do every day, which helps accelerate our path to being able to do the really cool AI stuff.”
Kenney also emphasized the need for creators of solutions to understand UX and to do so by testing using some of the same people that will ultimately use the solution. He said that is useful not only for AI but also for content creation and bringing the next generation into the workforce.
“When we think about where we are today, the digital natives that are coming into the U.S. government workforce,” Kenney said. “They’re expecting best of breed interfaces because that’s what they gravitate towards.”
“There’s a reason why people talk about the critical mass of adoption,” he added. “That when you’re a startup, you’ve got to get to that critical mass of users. And you don’t get to that critical mass of users in the civilian world if you have a crappy user interface and people don’t want to use it. … That’s exactly how people are looking at it in the government today.”
He mentioned the success of multivariate A/V testing that social media has tried to test certain features with its user base and said the Federal government should be thinking about testing user interfaces the same way.
“The creative process is also about discovering,” Kenney said. “You’ve got to be able to figure out where people are going to best be comfortable. We don’t always know that. The designers of the systems, you’ve got to let the users help you with them. But the government isn’t looking at it in that perspective across the board.”