Government leaders tout AI successes at Google Cloud Public Sector Summit

At the outset of the pandemic, public sector organizations saw large spikes in call center activity and website traffic, forcing them to quickly turn their attention to artificial intelligence-enabled tools to help with the initial pandemic response.

According to Marco Palermo, director of digital government and modernization for the City of Toronto, their initial AI pilot program has been so successful for the city that leaders are looking at other programs where they can expand the tool’s use.

“Getting information out to our citizens was extremely important. And we were looking for alternative channels to deliver that information,” Palermo recalled in December 2020 at the Google Cloud Public Sector Summit.

He described how city officials first reached out to their existing vendor partners for help. This is how they learned about a partnership between Cisco and Google and subsequently turned to Google Cloud’s Dialogflow, an interactive voice-response platform that uses natural language understanding to answer different questions citizens were asking.

Palermo said the chatbot setup was surprisingly quick. On April 1, 2020, they had their first conversation with Google and by April 4, they had a prototype in place. Then, between April and May, city employees worked to add content and make their AI chatbot smarter and interact with citizens more accurately.

The success of the platform surprised Palermo and Toronto leaders. He cited recent figures showing the chatbot answered 80% of questions.

In many ways, the pandemic amplified a need that had been bubbling beneath the surface for many public agencies, said co-panelist Franco Amalfi, strategic business executive for Google’s public sector operations.

“Even before the pandemic people were telling us that they want easy, flexible personal experiences with companies and as well as a public sector,” shared Amalfi. “People want support outside of the normal 9-to-5.”

Palermo’s story was part of a larger discussion on the advantages of AI in government. During Google’s Public Sector Summit, held virtually on Dec. 8-9, 2020, a number of government IT leaders discussed ways AI tools have helped to solve a variety of scaling challenges.

Improving medical outcomes for patients

In addition to addressing communications challenges, AI has also helped public agencies analyze complex large-scale, unstructured data repositories, like medical image files. That was the situation Dr. Hassan Tettah found himself in, trying to improve patient outcomes for the Department of Defense.

Using AI to identify and mitigate health challenges and provide better outcomes for patients is the primary driver for Tettah’s research. As health mission chief for the Joint Artificial Intelligence Center (JAIC), he is helping the DOD to advance predictive analytics in their diagnostic capability with certain cancers.

“In the DOD, cancer diagnosis treatment in the defense health agency alone costs over $1.7 billion,” he said. AI holds the potential to provide huge cost savings with early diagnosis and effective treatment options. He discussed the research in a panel, AI Fuels Smarter Medical Imaging and Better Outcomes.

Using an AI-augmented microscope, they have reduced time it takes to review a biopsy specimen of breast and prostate cancer by 30 seconds on average — or about a 12% savings in overall review time, Tettah said.

In order to train these algorithms, Google prototyped the new microscopic tool they call an “augmented reality microscope,” explained Dr. Craig Mermel, staff research scientist at Google. This tool brings AI analytics into the microscope.

The microscope uses a beam splitter to grab a copy of the image and passes it on to a high-resolution, high-speed digital camera. The camera passes those images on to an associated compute unit which analyzes the image and renders a prediction of, and where, the cancer might be present. Then the image is projected back to the field view, superimposing a form of augmented reality — an information layer — on top of the tissue image the pathologist is viewing.

“This entire process is happening incredibly quickly. We can go from an image back to a result displayed with oncologists and about 30 milliseconds,” explained Mermel. All of this happens in real-time as the slide changes, or the view magnifies.

“2020 has shown the maturity and power of applied AI to enable transformation, boost productivity and drive critical outcomes at scale. The power of AI to improve service delivery and to power public sector transformation will play out well into the future,” said Mike Daniels, vice president for Global Public Sector at Google Cloud, during the opening keynote at the summit.

While AI has proven helpful, public officials still face challenges — like fostering cultural acceptance of AI programs — which leaders said that their peers should consider as they build any new AI program.

Making strategic decisions to start an AI project

In the session All the Right Moves, Prioritizing Investments in Technology, government leaders recommend finding key use cases that show high value for business and a return on investment as a critical first step to AI adoption. According to Dominic Sale, assistant commissioner of Technology Transformation Services at the General Services Administration, getting employees comfortable with applying AI can be a struggle. One of the first critical steps in getting internal buy-in is to find key use cases that show high-value for the business and therefore on investment for AI adoption.

Sale also advised that before agencies begin a project, leaders need to look at institutional barriers preventing them from moving forward. These could be things such as a lack of talent needed to adopt and manage these tools or inadequate policies to address potential bias that may do more harm than good.

Former U.S. CIO Suzette Kent encouraged government leaders nevertheless to take a first step and get hands-on knowledge in applying AI through pilot programs.

“What’s really important is to get started, to understand the journey and begin to apply it to something that is specific to a mission,” said Kent. Pilot programs help leaders form broader strategies around sound IT investments and can help them future-proof mission delivery, she said.

Learn more about how Google Cloud is helping government agency leaders implement successful AI projects to improve mission outcomes.

View these sessions and more at the Google Cloud Public Sector Summit.