The Department of Homeland Security (DHS) is turning to responsible AI to implement human-centered design across the department to serve its many customers, according to DHS Chief Scientist Sam Howerton.
At the GovAI Summit on Tuesday in Arlington, Va., Howerton explained that DHS has one of the largest contact surfaces with people of any Federal agency. Its customers include travelers going through the Transportation Security Administration’s (TSA) airport screening, those engaging with the U.S. Customs and Border Protection (CBP), or even those whom the Federal Emergency Management Agency (FEMA) helps following a disaster.
“All of these people are our customers – and even more,” Howerton said. “So, what does the human-centered design approach look like for them? And what does it mean when we think about AI?”
“Responsibly-designed AI for the department gives us a chance to provide services at scale and speed that we have not been able to before,” he added. “It’s going to reduce the friction that people have when they work with the department.”
For example, Howerton said that an AI-design approach could help FEMA to make its grant process easier because “nobody wants to be filling out forms in triplicate following a natural disaster.”
Additionally, he said implementing generative AI at the U.S. Citizenship and Immigration Services (USCIS) can help those who are going through the immigration process and whose first language is not English.
“The future of homeland security, I think from a customer perspective, is one driven by decreased friction, faster response times, and let’s be candid, easier to understand information – because sometimes those in government like to use a lot of arcane words,” Howerton said.
“I would also say that – and maybe this is to be bold – the future of customer service, a human-centered DHS is actually one where people forget about us, where we just do our job and things work until they don’t and then we’re there to help you,” he said.
In addition to its customers, Howerton noted that DHS also needs to consider its operators across the country, including its colleagues in state, local, tribal, and territorial (SLTT) governments.
The chief scientist said this can pose “a really hard problem,” because the needs of each of these operational communities are different and ever-evolving.
“The underlying technology is the same, but the operational context is different,” he said. “So, our human-centered approach has to take into account not only our customers and our operators, but I would dare say the operators of the future as well.”
