Is Your Agency Prepared to Incorporate AI and AR?

Augmented Reality (AR) is more than just a customer-facing tool. It can also be used in the workplace to empower the federal workforce and improve the employee-customer experience. A session at IBM’s ThinkGov explored the opportunities created by AR and artificial intelligence (AI), potential difficulties with implementation, and case studies of successful projects.

The speakers at the panel were:

• Franz J Gayl, Science and Technology Advisor, Plans, Policies, and Operations Department, Headquarters Marine Corps
• Anil Tilbe, Director of Enterprise Measurement and Design Program, Veterans Experience Office, Veterans Affairs Department (VA)
• Lee Becker, Chief of Staff, Veterans Experience Office, VA

Kathleen Urbine, Partner, Mobile/Emerging Technology Lead, IBM Global Business Services, moderated the panel.

Urbine asked about the existing use cases that the speakers are starting to develop within their agencies.

Becker talked about the excitement around using AI and AR to improve operations. However, the agency is taking a step back because it realized that while it had been doing a lot of great things or things that it thought were great, the agency forgot to ask the veterans what they needed.

To reorient employees on the things that matter, the VA developed the very first journey map of the veterans’ experience. The Veterans Experience practice has a four-pronged focus:

1. Data

2. Developing tools

3. Technology

4. Engagement

“We’re trying to create an environment to allow AI to flourish,” Becker said. “The way to do that is to ask our customers and employees. That’s the journey that we’ve started.”

Tilbe emphasized the importance of establishing use cases with data. The VA has implemented suicide prevention programs.

Deploying a data science operation is crucial, Tilbe explained. Deep learning has a lot of integration potential when connected to AI and virtual reality (VR). One domain within data science is machine learning, which requires a lot of learning in terms of basing learning off of use cases. In the VA, for example, there is a lot of medical data in paper form. AI, deep learning and machine learning necessitates structured data, which posed a challenge. The technology experts have to take the data, scan it and process it by converting it to text. Then they have to analyze the text and employ AI capabilities.

“The architecture that is required to pull information has to be set up in a certain way,” Tilbe said.

In Gayl’s department within the Marine Corps, people embrace realism while also being enthusiastic and idealistic about the changes around them. “If AI and machine learning can help us, to the extent that it helps or improves upon the manual process that it replaces, we’ll adopt it,” Gayl said. “But we have to do so conservatively.”

The Marine Corps has more of an operational flavor and does not have a huge research budget. The staff watches to see where other groups are succeeding well in the Defense Department (DoD) and in industry. “We’re all about modernization, but modernization has to make sense,” Gayl stated. “We won’t adopt AI and AR for its own sake.” However, he added that the team is really excited about what has been accomplished so far.

Urban asked Tilbe how the VA is managing expectations around what is available today. He responded by stating that they’ve collected millions of data files in terms of qualitative input. There is a strong emphasis on the structured nature of data in the way the VA collects and analyzes data.

“The best way to reduce the margin of error is to ensure that the data is as structured as possible,” Tilbe said. “For example, using machine learning and deep learning, how do you predict where discrimination is taking place in the VA? A lot of what we’re seeing is aligned to the nature of the data.”

Urbine transitioned to the topic of workforce management and asked Becker about what he was doing to train the VA workforce.

Becker pointed to a report titled “More than Meets AI” that showed that the government has been through transformations in the past. “We’ve come a long way in terms of digitizing processes, but that took a lot of time,” Becker reflected. “It took an investment of our time to invest in different initiatives, and here we are in this new era.”

“‘More than Meets AI’ showed us that while 130,000 jobs are going away, new jobs will open up,” Becker continued. “We’ve been through this before, and this is something we have to tackle again.”

Tilbe highlighted the need for more data libraries that can easily be utilized. “We need more libraries so that we can automate more and code less,” he said.

Becker concluded by stating that considering the pains of government employees is of the utmost importance when implementing technologies, as well as working on building trust in those new technologies.