Interminably long on-line job applications are a horror for job-seekers. But it's not much better for hiring managers, who can only guess they are culling out the best job candidates.

Hirevue, Utah-based firm, has developed a machine-learning algorithm to appraise applications done by video. The system grades the applicants according to their suitability for the job, as set out by the hiring company. Human hiring managers can do the rest. "It allows hiring managers to not just randomly throw people out. You can spend time in ways uniquely human, rather than looking at resumes all day," said Lindsay Zuloaga, a data scientist with Hirevue.

As such systems improve, AI can take over more of the hiring process. Unilever, for example, is using online games and videos to weed out applicants before its human hiring people take a look at them. "It would be very cool if algorithm could get better and better at assessing people, and could hire someone," Zuloaga tells Axios."[A]n algorithm could say, 'This person is your best bet. They're going to stick around awhile,'" she said.