Over the past 20 years, government agencies have worked to shift from paper-based to digital operations. Early efforts, too often, went halfway. Agencies digitized the front end, but back-end operations largely remained the same. And on the front end, a digital customer experience was not as robust as it was in person. For example, early-stage digital government offered citizens access to forms online, which was convenient. But if someone had a question, there was no mechanism for in-person follow-up or an immediate response.
A series of emerging technologies – including natural language processing, artificial intelligence (AI) and machine learning (ML), new security and data management technologies, and 5G – are providing new opportunities to deliver a complete, more predictive, and overall better digital experiences for government employees and citizens accessing digital government services.
Natural language processing and natural language understanding are very important to mature digital government, says Marc Hamilton, vice president of solutions architecture and engineering at NVIDIA. “They enable agencies to move beyond digitizing data to turn data into knowledge, and knowledge into understanding.
“Think about this,” Hamilton adds. “If you could understand every call center recording from the last ten years in a particular agency and you fed that information into a recommender system, citizens calling in are going to get much more accurate recommendations than a simple keyword voice response system can offer.”
Digital-First Government Can Transform Services Delivery
A digital-first government is more connected, automated, data intensive, and distributed. It enables agencies to:
- Support the best possible workforce and citizen experiences
- Manage enormous growth in data volume, variety, and velocity
- Put emerging technology to work to make a positive impact on Federal missions
- Strengthen and modernize cyber defenses
Americans support the push for digital-first government. In a nationwide poll conducted by The Associated Press-NORC Center for Public Affairs Research and MeriTalk, 78 percent of more than 1,000 adults say increased access to government services online would impact them personally. And while just 9 percent say they are confident in the government’s ability to spend money in the best interest of the public, 68 percent say Federal spending on technology is worth the investment.
At the Department of Agriculture (USDA), a three-year partnership with General Service Administration (GSA)’s Technology Transformation Services Center of Excellence advanced digital-first government via projects focused on cloud adoption, contact centers, customer experience, infrastructure optimization, and data and analytics.
Together, USDA and GSA were able to:
- Close 31 of USDA’s 37 data centers, saving the agency $42.5 million
- Modernize the USDA website
- Consolidate 12 USDA online contact centers and launch AskUSDA, which serves as the front door for 15 USDA contact centers
- Save $10 million by reducing data analytics duplication
- Launch a central USDA cloud office
- Train 3,000 USDA employees in using dashboards and data visualization; and
- Increase USDA Farm Loan approval rates by 3.5 percent
“This project is a prime example of GSA’s commitment to using technology to help Federal agencies deliver for the people and communities they serve,” says GSA Administrator Robin Carnahan.
AI Anchors Digital-First Government
Operationalizing Federal AI is the cornerstone of digital-first government, according to 87 percent of 150 Federal IT decision makers surveyed in April 2021 by MeriTalk, in partnership with Dell Technologies and NVIDIA.
But the majority also say their agency is struggling to incorporate AI pilots or skunkworks programs and into overall IT operations.
To get over that hurdle, agencies must first define the specific outcome they want the AI project to achieve and then work backward toward an IT plan that will make that outcome possible, advises Jay Boisseau, artificial intelligence (AI) and high-performance computing technology strategist at Dell Technologies.
“It is so important to have someone on your staff or a service provider who truly understands your domain and your mission problem, and also understands data science and how IT capabilities map to data science capabilities,” he says.
Once the agency defines the data science capabilities that the AI project requires, the agency can build an IT plan to deliver those capabilities. The plan should encompass, edge, core, and cloud, so that the agency can optimize costs and resiliency, Boisseau says.
Flexibility is key in designing for digital-first government, he notes.
“Applications need to be developed and run-on infrastructure that is designed to allow for flexibility and scalability,” Boisseau says. “In a digital-first government, back-end applications are not tied to specific locations and infrastructure, but instead employ cloud-native principles and virtualization technologies so that agencies can put the right workload in the right place, and resiliency is built in.”
IT plans may include data processing units, a new class of programmable processors. Data processing units (DPUs) promise to accelerate data processing and improve performance in a host of other areas, Hamilton noted. A DPU is a system on a chip that combines:
- A high-performance, software-programmable, multi-core central processing unit (CPU), which is tightly coupled to the other security operation center (SoC) components;
- A high-performance network interface capable of parsing, processing, and efficiently transferring data to graphics processing units (GPUs) and CPUs at the speed of the rest of the network; and
- Flexible and programmable acceleration engines that offload and improve application performance for AI and machine learning, security, telecommunications, and storage, among others
Smashing Data Silos – and Security – Helps Agencies Put Data to Work
Comprehensive data management is also an essential building block of digital-first government. To take advantage of rich data analytics techniques such as natural language processing, agencies need to leverage increasing volumes of data, which requires breaking down data silos between organizations.
“You want to open up internal silos so that you can empower your data scientists to do things they could never have done with the previous generation of data analytics techniques and tools,” Boisseau says.
Data strategies, AI, and machine learning are making that possible, says Gundeep Ahluwalia, CIO of the Department of Labor. In the past, “When Congress wrote a law … we would respond to it by creating a fragmented piece of data that was fit for one purpose and one purpose only,” he says. “The more fragmented we got over the years, the harder it becomes to bring us together. [But] I think there is enough focus on data strategies now, and the technology has evolved with AI, ML, and other tools allow us to reintegrate the data that was so fragmented.”
In addition, security must be baked in, not bolted on, to adequately protect agency networks and the increasing volumes of data that are leveraged to enable digital-first government.
“Chief data officers and chief information security officers have an interesting challenge: As AI and other emerging technologies are increasing the value of data inside the agency, the data is becoming increasingly valuable to competitors and bad actors,” Boisseau says. “Security must pervade every aspect of the technology ecosystem, from hardware manufacturing and software development to solution integration, the supply chain, and employee training.”
The majority of government cyber leaders echo Boisseau, according to a recent survey of more than 300 cybersecurity leaders across Federal, state, and local government. The survey found that 83 percent of public sector organizations operate on an “assume breach” model today. And, 70 percent estimated their high-value assets (HVAs) have been compromised in the past 12 months. As a result, 91 percent say they want to see their organization shift from “assume breach” to breach prevention in the next three years.
While almost all respondents reported taking steps to improve risk management, just half say they are making progress on foundational cyber hygiene, including enforcing multifactor authentication and encryption, deploying endpoint detection and response systems, and auditing hardware security. Only 45 percent of organizations have developed a prioritized list of HVAs.
Eighty-nine percent of government cyber leaders say further prioritizing platform security will help achieve improved breach prevention. It will help organizations improve their ability to isolate critical infrastructure from vulnerable devices, as well as reduce exposure and risk.
To improve cyber breach prevention, government cyber leaders say they need:
- Centralized access to cybersecurity data and analytics (91 percent)
- Improved vulnerability management (90 percent)
- Hardened endpoint devices (89 percent)
- Fundamental culture change (89 percent)
- Increased investments in zero vulnerability solutions (89 percent)
“Cyber leaders are underwater, but it is possible that we can move toward a reality where breaches are not a given,” says Jimmy Sorrells, president, INTEGRITY Global Security.
“The industry needs to know that there are zero vulnerability platforms available, and those platforms are the key to helping our public servants better protect critical systems and citizens,” says Sorrells. “It is going to take a stronger commitment to cyber hygiene, platform security, and breach prevention to make real progress. We cannot continue to do the same things and expect different results.”
5G Will Amplify Digital-First Government Capabilities
IT plans should also take 5G networks into account because it will enable agencies to reap even more value from their data as they build digital first government services. In fact, Anne Neuberger, deputy national security advisor for cyber and emerging technology, calls the power of 5G transformative.
“It will touch every aspect of our lives … underpin critical infrastructure such as transportation and electricity distribution, and enable advances [in] autonomous vehicles and remote healthcare, and perhaps other applications we can’t even think of,” she says.
5G will enable widespread computing at the network edge, and it will enable agencies to process 500 times or more data than today.
“Ultimately, with 5G more compute will be done at the edge than in all data centers today,” Hamilton predicts.
To successfully leverage 5G and AI at the edge, Hamilton advises agencies to implement a platform that offers centralized management across the entire infrastructure.
“Agencies that are implementing their first AI projects into production in their data center don’t need the burden of learning a new hardware platform for deploying to the edge with 5G,” he says. “They need a platform that scales from the data center to the edge to the cloud so that they have consistent, streamlined operations.”
Modernization Will Fuel Continued Progress
The Federal government is making progress in its digital-first strategy to improve the delivery of citizen services, and it must continue to adopt new technologies, new techniques, and new approaches to keep up with the pace of change, notes Deputy Federal CIO Maria Roat.
“Digital delivery is a necessity, and our customers – internal employees [and] the external American public – expect that delivery to be quick, easy, secure, and … accessible. It’s a business discipline, we must embed it in everything we do,” she says.