GSA Blog

GSA focuses on transparent, equitable AI

When the first railway opened in 1825, people worried the initial speed of 30 mph might melt them. In the 1920s, it took some convincing for people to embrace electricity. As recently as 1993, a survey reported that about 2.5 million people in Great Britain feared using the telephone. 

Artificial intelligence is the most recent technology to create apprehension in some people, fueled by the powerful capabilities and potential use cases of this emerging technology.

“We’ve been here before. Remember that GSA was one of the first agencies to introduce things like the internet and email into the federal work environment,” says Zach Whitman, Chief AI Officer. 

“We didn’t know exactly how those tools were going to help us achieve our mission, but today they’re key parts of how we operate.”

AI is not as new as people think

The fact is, agencies already use AI to analyze weather hazards, process veteran feedback, answer questions about retirement, match skills to job listings, search patents and more. 

At GSA, AI is being used to route IT service tickets, that were manually sorted before, to the correct response team. And we’re finding new applications each day, such as AI helping draft renderings of potential new floor plans to support our real estate mission.

“We have been helping folks to get familiar with AI, including tools like ChatGPT and Google Gemini,” Whitman said. “And moving forward, our goals are to use AI to empower our workforce, making tasks easier, as we have in the past with other technologies, and ultimately delivering better services to our customers and the public. 

“The efficiencies we’re exploring include general applications like knowledge retrieval, predictive analytics and software development as well as more specific technical applications,” he said.

Guiding agencies on AI

There’s a whole external side to this for GSA, Whitman points out. 

“I’m just one of many Chief AI Officers throughout government and we are all working collaboratively to ensure that we build upon each other’s strengths, and GSA is helping lead the charge.”  

GSA’s Technology Transformation Services is an example of that. 

“TTS is helping lead the way for more and more agencies that want to harness the power of AI,” said Ann Lewis, TTS director. 

One TTS effort is the AI Community of Practice, which brings together thousands of government workers who are interested in helping their agencies responsibly use AI to deliver on their missions. This past fall, in partnership with Stanford University, the AI CoP hosted a six-part training program on AI for federal employees, and they expect to launch a new iteration of that training soon. 

The agency is also prioritizing cloud-based AI products through the Federal Risk and Authorization Management Program, or FedRAMP, which ensures cloud services used by the government are secure. Agencies are required to use only FedRAMP-approved technology.

Another example is the Generative AI Acquisition Guide to help the acquisition community across government in four areas: 

• Defining agency needs and utilizing AI to find solutions.
• Providing a test space to try solutions before committing to large-scale buys.
• Sourcing sound data inputs to ensure quality data outputs, while managing its use and protection.
• Monitoring and controlling costs. 

This new guide was created by GSA’s Federal Acquisition Service’s Office of Information Technology Category. Assistant commissioner Laura Stanton called for agency chief AI officers “to set up the right safeguards for how the AI tools your teams and others within your agencies use will meet cybersecurity standards and best practices.”

Overall, Whitman said GSA is working to help ensure tech is used appropriately and in support of equity and accessibility. 

Empowering workforces

As other innovations like email and the Internet have in the past, AI may require some training and evolution of the workforce. And like the telephone, trains and electricity, AI has the potential to propel innovation in government operations and services.

Whitman explains that quality AI will depend on programming input from humans who understand the goals to be achieved. This includes creating AI that supports the human element, such as conducting repetitive, non-skilled tasks in a fraction of the time and freeing humans up to be more productive in other areas, such as working with customers, developing higher-level strategies and conducting other non-repetitive tasks.

“Our vision is for AI to support the functions of our people,” Whitman says. “AI can free up some of the time we spend on repetitive and mundane tasks, allowing us to focus on the more critical aspects of the work. At the end of the day, it helps us deliver the best services to our customers, which is what it’s all about.”

GSA and agency teams will continue to be the drivers of the best decisions across our mission areas, he added. 

In March, the White House issued the first governmentwide policy on AI, directing agencies to “advance AI governance and innovation” while managing risks, particularly those that affect the public’s rights and safety. Managing risks includes strengthening AI governance and advancing responsible AI innovation.

Agencies are directed to use AI to: 

• Improve accessibility of government services.
• Improve public health and public safety while ensuring privacy.
• Protect democracy and human rights.
• Grow economic competitiveness to benefit people across the United States.
• Enhance accessibility and efficiency in customer service, planning and budgeting, and employee experience.
• Improve acquisition through more organized solicitation reviews, evaluations and data transactions.

Next steps

Agencies considering or using AI need to continually assess AI’s potential impact on equity, fairness, accessibility and more. Getting feedback from the public and affected communities will be important and GSA invites the public to share their thoughts. 

Meanwhile, GSA and other agencies will continue to build AI expertise in government.  Already, there has been a 288% increase in hiring for AI-related work roles, with growing interest in programs like GSA’s Presidential Innovation Fellows and the U.S. Digital Corps.

To support more hiring of tech workers, GSA’s Technology Transformation Services joined Tech to Gov career fairs that resulted in the resumes of thousands of qualified applicants, leading to more than 100 interviews and 62 candidate selections. Forty are already at work in TTS.  

If you’re already in government, “a good place to start is the AI Community of Practice because that’s how we’re building capacity across every level of government,” Whitman said. “It’s the go-to place for practical implementation of responsible AI across government — and it has over 4000 members from GSA, Homeland Security, DoD and elsewhere.