For decades, much of the federal government security issuance process relied on techniques that emerged in the mid-twentieth century.
“This is very instructive,” said Evan Lesser, president of ClearanceJobs. “Walking by car to meet people. It is very old and takes a lot of time. “
The federal initiative, launched in 2018, called Trusted Workforce 2.0, has officially introduced the semi-automated analysis of federal employees that takes place in real time. This program will enable the government to use artificial intelligence for employees who are seeking or already have a security clearance, under “continuous review and evaluation”. .
“Can we build a system that checks on someone and continues to check on them and understands that person’s mood because they exist in legal systems and public record systems on a continuous basis?” Said Chris Grizhalva, chief technical officer of Peraton, a company focused on the government side of insider analysis. “And from this idea came the notion of continuous evaluation.”
Such attempts have been used in government in more ad hoc ways since the 1980s. But the 2018 announcement was aimed at modernizing government policy, which typically evaluated employees every five or 10 years. The motivation for policy and practice adjustments was, in part, to pass the necessary investigations and the idea that circumstances and people are changing.
“That’s why it’s so persuasive for people to be under a constant, ever-evolving surveillance process,” said Martha Louise Deutscher, author of System Screening: Identifying Security Threats. He added: “Every day you conduct a credit check, and every day you run a criminal check – and bank accounts, marital status – and make sure that people do not fall into circumstances where they can. Would become a risk if they were not yesterday. ”
The first phase of the program, the transition period to full implementation, was completed in the fall of 2021. In December, the U.S. Government Accountability Office recommended evaluating the effectiveness of automation (though not, you know, consistently).