How a Robot Could Score Your Likelihood of Obtaining a Security Clearance

A RFI from Immigrations and Customs Enforcement (ICE) is looking for a system to provide scoring for security clearance applicants. It’s like the government’s electronic adjudication program, applied on the front end of the security clearance process.

The use of the SF-86 as an employment screening tool by private contractors has been discouraged, because of the risks of OFCCP violations or lawsuits – the demographic information included in the SF-86 could be seen as used to discriminate, which would create issues for cleared employers

I’ve been asked to fill out a pre-employment security screen; not exactly an SF-86 but similar in scope. Is that the same thing? I think I’ve read posts on here from people who did have to fill out an SF-86 during the employment pre-screen.

Regarding the automated scoring system, I wonder how they would “train” the system. In today’s “Big Data” world they could conceivably dump a ton of completed SF-86s in there with associated results, but I think that would violate a lot of rules. So probably they will have to get some human “experts” to train the system.

My father has been a county sheriff for 40 years and is looking to get a part time job as a background investigator. He most recently ran the internal affairs department. I am a commander in the military with 29 years of service and gave him the idea because I have been interviewed by civilian investigators many times. However, I have no idea what the companies are or how to apply. I was wondering if anybody out there could help. Thanks!

If you search for Field Investigator with Securitas Critical Infrastructure Services, Inc you’ll find some positions.

If you are not already credentialed, the likelihood of you becoming an investigator is slim to none. Companies are not holding new hire orientation

Adding to that, if you don’t have a year or more of experience conducting background investigations, don’t bother.

I would also be concerned about potential biases the AI may have. Time and again we’ve seen predictive algorithms reproduce the same biases their human programmers have (unconsciously or otherwise).