Vetting Afghan recruits faster without compromising personnel safety

Insider or "green-on-blue" attacks by hostile Afghan forces seeking U.S. training or employment were a deadly threat to U.S. personnel in Afghanistan. Investigations highlighted the failure of existing vetting processes to reliably screen Afghan partner forces, making the reduction of human-based risk paramount.

SOCOM needed to vet hundreds of Afghan commando recruits—civilians and existing soldiers—to expand the country's defense capabilities without jeopardizing U.S. safety. The enrollment process failed due to a lack of immediate, actionable insider threat data and high-risk applicants defeating the system.

The most actionable data came from DNA comparison against known IED manufacturers and cellphone analysis. This analysis took 10 days to two weeks. During this wait, training continued, with recruits issued rifles and live ammunition, creating many opportunities for insider attacks.

To make the job of counterintelligence screening easier and more effective, SOCOM searched for an insider vetting process that:

Voice-based risk assessment to screen recruits

SOCOM selected Clearspeed to screen its commando candidates by leveraging its simple yes/no questionnaire. Questionnaires were conducted in tandem with the established bio-enrollment process that included personal biographical information, DNA swab, photograph, iris scan, fingerprinting, and various technological exploitation methods.

The team screened 715 recruits in just 20 hours, in both the Dari and Pashto local languages. The results (which were unbiased - age, ethnicity, language, military rank and unit, and place of residence did not influence outcomes) were used as an additional data source to SOCOM’s initial and on-going insider threat vetting processes. People flagged as low or average risk could be cleared through the process much faster, unless there were other indicators to the contrary. Individuals with a high-risk level may have required additional follow up based on the question and the risk profile of SOCOM and the role of the commandos. This allowed SOCOM to focus their counterintelligence resources on the areas of highest risk.

Clearspeed identified 100% of those commandos flagged as ‘at risk’ by the existing process. In addition, Clearspeed identified eight candidates who were confirmed as ‘high risk’ but not identified by the current process, and three candidates who, upon notification of being flagged for risk, went AWOL — indicating the potential that they were not identified as ‘high risk’ by the current process. During this project, Clearspeed demonstrated a >97 percent accuracy rate, that was all unbiased, and presented no false negatives.