When the office snitch is a machine


Workplace surveillance has been a controversial subject for as long as there have been workplaces. Advocates of surveillance insist that it is essential for productivity, quality control, and (in some cases) safety. Further, they’ll insist, employees have no inherent right to privacy at work.

Opponents of workplace surveillance argue that the practice actually is counter-productive; employees feel as though they aren’t trusted, which lowers their morale and, inevitably, their effectiveness. Some also raise questions of legality.

Technologies such as video cameras, electronic ID badges, and keylogging software have made it easier for employers to track workers’ locations, activities, and physical output. Now, thanks to artificial intelligence (AI) and machine learning, employers may be inching closer to getting into workers’ heads — and that has dangerous implications.

In this Fox News article on using AI to monitor workers, it becomes clear that much of what AI is doing is just a more efficient version of what legacy workplace surveillance technologies already are doing — logging activity, measuring output and outcomes, and providing a basis for calculating the value of individual employees.

But AI also can take things to another level:

Matt Fleckerstein, a spokesperson for workflow company Nintex, told Fox News that many companies already track employee browsing habits. …Where AI can help is to determine the intent. Someone might be browsing articles all day, but there might be a good reason for that — such as conducting research.

I know this may seem like an innocuous example, but inherent in the notion of “intent” is an assumption about a person’s state of mind. Who’s comfortable with that as a metric for gauging an employee’s effectiveness or level of commitment to the job?

Probably not Jacob Silverman, author of Terms of Service: Social Media and the Price of Constant Connection. In a lengthy critical analysis of wildly popular collaboration platform Slack, Silverman writes that “Slack slots neatly in the trend toward the gamification of labor and everyday communication.”

But something more is going on underneath the hood. Slack “tracks and catalogs everything that passes through it,” Silverman writes, creating a Searchable Log of All Conversation and Knowledge (the company’s name is an acronym derived from this!) available not just to team members, but to anyone with access to the data:

Forget the annual performance review; with Slack’s help, managers could track their employees even more closely, and in ever more granular ways. And why stop at performance analytics? Sentiment analysis could automatically alert supervisors when employees’ idle bickering tips into mutiny. Depressed or anxious employees could be automatically served with puppy videos and advice bots.

Or they could be “downsized” if someone determines their perceived psychological issues may negatively impact future performance and productivity.

Am I being too paranoid about this, or is the marriage of increasingly sophisticated surveillance technology with AI, machine learning, cognitive analytics, and predictive analytics a potentially dangerous development? Let us know in the comments section below.


  1. […] in fact! Four months later I wrote about how AI can be used to interpret employee actions (such as web browsing) to discern “intent,” and to conduct sentiment analysis on […]

Speak Your Mind


This site uses Akismet to reduce spam. Learn how your comment data is processed.