IBM has developed artificial intelligence that can predict, with a reported 95 percent accuracy, when employees are just about to quit their jobs.

Super snoopy surveillance? Or smart technology that can actually benefit both business and employee? The jury’s still out. It’s all in the ultimate application.

The justification for the “predictive attrition program,” as IBM dubbed it on its pending patent, is to determine when employees are poised to quit in order to see if something can be done to stop them. Retention rates, after all, translate into big bucks for businesses. IBM’s chief executive Ginni Rometty estimated the program’s already saved $300 million in her company’s “retention costs,” CNBC reported.

There’s no real fool-proof way to verify those savings.

But revolving doors and employee departures are well-known, well-documented costly managerial headaches. So it’s true, technology that can give company chiefs a heads-up when their well-trained, well-regarded workers are about to leave — and therefore, a chance to swoop in with offers of promotion or financial and educational benefits — is a boon for the business. It’s also a potential boon for the employee.

The company may save on the costs and frustrations of rehiring and retraining, but employees could actually receive counter-offers that prove more lucrative than the ones that were enticing them to leave.

Win-win. Right?

Still, IBM isn’t exactly releasing details on how its A.I. works. And therein lies the warning.

Companies already, with alarming regularity, monitor employees via hidden cameras, technology in the copy machines, software that records keystrokes, computer programming that reads emails and watches Internet surfing — and more. The justification? To curb wasted manpower hours, stop employee thefts and protect corporate higher-ups from surprise legal entanglements.

So it’s somewhat unnerving to add this new A.I. level of employee monitoring to the employer’s list.

If IBM derives its information based on information employees give during, say, performance reviews or at company meetings, that’s one thing. That’s pretty above board and hardly surprising. But what if some of the information comes via facial reading technology, where the A.I. software analyzes the images of employees regularly captured on the company’s cameras?

Or, from iris scanning at the door that determines the level of employees’ excitement as they head into the office?

That’s not just unsettling. That’s intrusive. It’s also a recipe for massive misinterpretations of data and misunderstandings of employees, leading to grievances that press workers to find new jobs, anyway — and the very thing the company was trying to prevent coming true.

Besides, don’t successful managers know already how their employees feel about their jobs? If not, they should. If a company’s only as good as its employees, well then, executives shouldn’t be so far removed from their bread and butter, so to speak, that they have to rely on machinery to tell of impending departures. They should intuitively know; they should instinctively realize.

Again, it’s too soon to tell about this new technology. The proof is all in the applications, the potential ranges from win-win to lose-lose.

But one thing’s clear. The days of burning off those resumes on the company printer on the sly are definitely at an end.

First appeared at The Washington Times.