Support for military families

  • Project: Supporting Military Family Cybersecurity with an AI-Infused OSINT Tool
  • Principal investigator: Kurt Luther, Department of Computer Science and Institute for Advanced Computing
  • Co-principal investigator: Jin-Hee Cho, Department of Computer Science

Service members and their families can be major targets for cyberattackers, including from foreign adversaries. Their knowledge of sensitive information, such as troop deployment locations and schedules, and even their access to government benefits and pensions can be attractive to cyber criminals.

“We know from previous work that there is public information about military members that can be used for cyberattacks, including information found on social media accounts,” said Luther, whose areas of expertise include human-computer interaction and human-artificial intelligence (AI) collaboration. “Less attention has been paid to the family members so that is the angle we are taking.”

The research led by Luther and Cho aims to safeguard U.S. military families from adversaries who scour public websites for operational clues. The one-year study started this summer with interviews of military family members and unit security managers to understand how they currently manage the ongoing challenge of operational security (OPSEC).

“Generally, military families are doing a good job practicing OPSEC,” Luther said. “But there are ways that information can leak out, even inadvertently or through aggregation, and that’s the problem we are looking to solve.”

Cho has previously conducted research to assess security risk for vulnerable populations such as seniors, and she and Luther plan to adopt that methodology to the subject of military families. This summer’s interview findings will guide construction of a new Operational Security Cyber Vulnerability Index, a quantitative score combining items that contribute to higher security risk such as having a large number of active social media accounts in the family or living in geographic areas with increased adversary capability.

The project has four phases with phases one and two focused on development of the vulnerability index. Then in phase three, the team will integrate this index into a software prototype, Homefront Helper, that silently inspects social-media posts, flags text or images disclosing potentially sensitive information, and offers clear recommendations that users may accept or refine.

“The idea is to develop a research-backed tool to promote community learning and reduce the likelihood of inadvertent OPSEC violations,” Luther said.

The project also has support from the Boeing Center for Veteran Transition and Military Families based at Virginia Tech in the D.C. area. The center, through Director Debbie Bradbard, is connecting the Virginia Tech team to local military leaders and families in the area.

“Military families use social media regularly, but they may only have a limited understanding of how their online behavior impacts OPSEC for their service member,” said Bradbard. “This study will help provide tools to military community to help change behavior and avoid unintended sharing of security information.”