The House Subcommittee on Cybersecurity, Information Technology, and Government Innovation sunk its teeth into robotic dogs, drones, and other technologies it identified as important for law enforcement and military applications at a June 23 subcommittee hearing.

The hearing featured uses cases for those technologies, along with videos and demonstrations of drones and robotic dogs that are beginning to be employed by Federal agencies.

“We’re going to explore the use of cutting-edge technology to protect America’s citizens and its borders,” said Rep. Nancy Mace, R-S.C., who chairs the subcommittee. “The innovations we’ll discuss also help to protect border security officials, law enforcement officers and military personnel who put their lives on the line each and every day to keep the rest of us and our nation safe.”

Prominent during the hearing was discussion of use cases for deployment of AI technologies on the U.S. southern border.

Ryan Rawding, vice president of business development at Pangiam, discussed how AI technology is being used by the U.S. Customs and Border Protection (CBP) to better identify migrants crossing the southern border.

“CBP collaborated with Pangiam to develop the mobile intake app … Mobile intake utilizes AI and computer vision to capture the subject’s face and leverages optical character recognition to extract biographical information from documents and track migrant’s properties,” said Rawding.

Although AI has shown to be helpful in migrant documentation Rep. Mace raised concerns about adverse impacts of new tech deployment in the future.

“Thinking about some of the concerns – whether it’s data security, cyber security – the advent of AI is moving technology and cybersecurity threats in rapid pace … how do we prepare ourselves for that future,” Rep. Mace asked. “What are some of your concerns? How are we addressing those concerns?” she asked.

Rawding responded by emphasizing that humans need to remain in the loop with AI tech.

“In my experience, employing a lot of supervisory machine learning models in the customs and border protection space, the computers and machines are not ultimately making the decision but giving risk-based scores,” he said, adding that agents and officers are then able to decide whether or not to take action based on that data.

Read More About
More Topics
Jose Rascon
Jose Rascon
Jose Rascon is a MeriTalk Staff Reporter covering the intersection of government and technology.