[ad_1]
John P. Desmond, Editor of AI Trends
AI gives them experience-based expectations of what is more accessible and possible for young people in the workforce who grew up as Alexa and self-driving cars as part of the landscape.
That idea set the foundation for the panel discussion at AI World Government The concept needs and skill set myth of the AI engineering team held virtually and in-person this week in Alexandria, Virginia.
“People feel that AI is within their grasp as technology is available, but this technology is ahead of our cultural maturity,” said panel member Dorothy Aronson, CIO and Chief Data Officer of the National Science Foundation. “It’s like giving a child sharp objects. You might have access to big data, but that might not be the right thing to do.”
Things are accelerating and raising expectations. When panel member Vivek Rao, a lecturer and researcher at the University of California, Berkeley, was working on his PhD, a dissertation on natural language processing could be a master’s thesis. “Now we assign homework on a two-day turnaround. We have a huge amount of calculation skills that we couldn’t use even two years ago,” he said of the students.
Panel moderator Rachel Dzombak leads digital transformation Software Engineering Research Institute At Carnegie Mellon University, we asked panelists what is unique about working on government AI.
Aronson said the government couldn’t move technology too far, or that users wouldn’t know how to interact with it. “We don’t have an iPhone,” she said. “We are experimenting and always looking ahead and anticipating the future, so we can make the most cost-effective decisions. Now, the government is seeing the convergence of emerging and retired generations.
Early in her career, Aronson didn’t want to work for government. “I thought it meant you were in either Armed Service or the Peace Corps,” she said. “But after a while I learned that federal employees are motivated to be a service to the larger problem-solving agency. We solve the big issues of equity and diversity, get food for people, keep people safe. People who work for the government are dedicated to those missions.”
She mentioned two children in her 20s, who like the idea of service, but in “a small chunk,” she says, “they don’t see government as a place where they have freedom, they can do whatever they want.
Berkeley students learn about the role of government in disaster response
Rao from Berkeley said his students were watching wildfires in California and asking who is taking on the challenge of doing something about them. When he told them it was almost always local, state, federal agencies. “Students are generally surprised to find it.”
As an example, he worked with the CMU and the Department of Defense, the Army Futures Research Institute, and the Coast Guard to develop courses on disaster response innovation. “This was eye-opening for students,” he said. Initially, two of the 35 students expressed interest in federal careers. By the end of the course, 10 out of 35 students had expressed interest. One of them was hired as a software engineer at the Naval Water Surface War Center, outside of Corona, California, Rao said.
Aronson describes the process of calling new federal employees “heavy lifts,” suggesting that “if they’re ready in advance, they’ll move much faster.”
Dzombak asked which skill sets and mindsets are considered essential to AI engineering teams. Panel member Bryan Lane, who stated that he was director of General Services Administration, who announced during the session that he was taking on a new role at FDIC.
Lane is a technology executive within the GSA IT Modernization Center of Excellence (COE) and has led senior analytics and technology initiatives for over 15 years. He is leading a GSA partnership with the DOD Joint Artificial Intelligence Center (JAIC). [Ed. Note: Known as “the Jake.”] Lane is also the founder Data XD. He also has industry experience and manages the acquisition portfolio.
“The most important thing about resilient teams going on their AI journey is to have unexpected preparations and the mission continues,” he said. “If you all agree on the importance of the mission, the team can hold it together.”
A good sign that team members are acknowledging that they have never done this before
As for his thinking, he said that many of his team members came to him and said, “I’ve never done this.” He sees it as a good indication that provides an opportunity to talk about risks and alternative solutions. “When your team has the psychological safety that they don’t know anything,” Lane considers it positive. “The focus is always on what you did and what you provided. It’s rarely focused on what you haven’t done before and what you want to grow,” he said.
Aronson finds AI projects difficult to move off the ground. “It’s hard to tell management that there’s a use case or problem that you want to solve it and go to it. And the chances that it will accomplish are 50-50 and you don’t know how much it will cost,” she said. “It will clarify the rationale and convince others that it’s the right thing to do to move forward.”
Rao said he was talking to students about experimental and experimental thinking. “AI tools are easy to access, but they can hide the challenges you can encounter. For example, applying the Vision API in the context of business or government issues, things may not be smooth,” he said.
Moderator Dzombak asked the panelists how to build their team. The arson said, “You need a mix of people.” She attempted a “community of practice” to solve specific problems that people could come and go. “You bring people together not around tools, you’re focusing on issues,” she said.
Lane supported this. “I stopped focusing on tools in general,” he said. He conducted experiments at JAIC in accounting, finance and other fields. “We found it wasn’t really about tools. It’s about bringing together the right people to understand the problem and then looking at the tools available,” he said.
Lane said he would set up a “sensual team” that was “a little more formal than the community of interest.” He found that they were effective in tackling problems for probably 45 days. He also likes to work with customers of the services they need within the organization, resulting in their customers learning about data management and AI. “We’ll pick up one or two on our way to advocate for accelerating AI across our organization,” Lane said.
Lane believes it will take five years to solve the proven mindset, work and best practices for developing AI systems to serve governments. He mentioned Opportunity Project (TOP) The US Census Bureau, which launched in 2016, is tackling issues such as marine plastic pollution, Covid-19 economic recovery, and disaster response. Top was engaged in over 135 public projects at the time, with over 1,300 alumni including developers, designers, community leaders, data and policy experts, students and government agencies.
“It’s based on how you think and how you organize your work,” Lane said. “We need to expand our delivery model, but in five years, there’s enough proof of concept to know what works and what doesn’t.”
See more details at AI World Government,in Software Engineering Research Institutein DataXD And Opportunity Project.