AI tracks your every move and tells your boss if you’re slacking
By Timothy Revell
30 January 2017
www.newscientist.com
Think government surveillance is excessive? Wait until you hear what your employer might be up to.
Companies are increasingly using technology to monitor employees in the workplace, with artificial intelligence making it possible to track individuals’ behaviour in great detail. Start to slack off, or show signs of going rogue, and an algorithm could tattle to your boss.
One company offering such services is London-based start-up StatusToday. The company was recently included in a cybersecurity accelerator run by British intelligence agency GCHQ, which offers technical expertise and helps to secure investment. StatusToday’s AI platform relies on a regular supply of employee metadata, including everything from the files you access and how often you look at them to when you use a key card at a company door.
The AI uses this metadata to build up a picture of how companies, departments and individual employees normally function, and flag anomalies in people’s behaviour in real time. The idea is that it could detect when someone might pose a security risk by stepping outside of their usual behavioural patterns. “All of this gives us a fingerprint of a user, so if we think the fingerprint doesn’t match, we raise an alert,” says Mircea Dumitrescu, the company’s chief technology officer.
The system could flag if an employee starts copying large numbers of files they don’t normally view, for example. Are they just going about their job, or could they be stealing confidential information? It also aims to catch employee actions that could accidentally cause a security breach, like responding to a phishing email or opening an attachment laden with malware. “We’re not monitoring if your computer has a virus,” says Dumitrescu. “We’re monitoring human behaviour.”
But catching the odd security breach in this way means monitoring everyone. “It seems like they’re just using the aura around AI to give an air of legitimacy to good old-fashioned work place surveillance,” says Javier Ruiz Diaz, policy director at digital campaigning organisation the Open Rights Group. “You have a right to privacy and you shouldn’t be expected to give that up at work.”
AI over your shoulder
Some companies already keep employee metadata for retrospective analysis if something goes wrong. Insurance firm Hiscox recently started using the StatusToday platform and immediately detected activity on an account from an employee who had left the company months earlier.
In addition to flagging potential cybersecurity alerts, the AI can be used to track employee productivity. Dumitrescu cites the example of Yahoo controversially banning staff from working from home on the basis that this reduced the “speed and quality” of work across the board. “We can actually quantify if this is true for individual employees,” he says. “Whether they should be allowed to work from home can then be based on data.”
Exactly how companies use the system will be up to them, but it’s hard to shake the image of an AI constantly peering over employees’ shoulders. “The privacy concerns with this kind of technology come down to consent and understanding,” says Paul Bernal at the University of East Anglia. “Are the employees aware and have they been given a choice?”
Monitoring people in this way could have adverse effects by making them change how they work, says Bernal. “The general creepiness will bother people, and that could be counterproductive if it affects their behaviour.”
It’s for the same reason that Phil Legg at the University of the West of England says flagging unusual behaviour will never catch every security risk. “If people know they’re being monitored, they can change their behaviour to game the system,” he says. With enough determination, a disgruntled employee could collect a set of damaging files one-by-one over the course of a year, potentially going unnoticed.
