IMAGE: Mohamed Hassan — Pxhere (CC0)

Algorithms now make it easier than ever for employers to build up a picture of you based on your social network profiles

Algorithms now make it easier than ever for employers to build up a picture of you based on your social network profiles

US startup Predictim has come up with an algorithm for families looking for babysitters or home help that calculates candidates’ suitability based on their social network activity, using metrics that aim to predict behavior ranging from disrespectful or negative attitudes to drug use, explicit language or harassment and bullying. After hiring somebody based on the result of the analysis, families can continue monitoring their new employee.

Other companies, such as Fama, algorithmically analyze the attitudes of a company’s employees in search of potential reputational risks also based on their social network profile, while HireVue does the job once carried out by psychologists by analyzing video interviews of job candidates using algorithms to identify tone-of-voice patterns, vocabulary or facial expressions and then predict their possible performance for a particular job.

There’s nothing new about behavior analysis, and I’ve discussed it on numerous occasions; the breakthrough here is to do it algorithmically, with all that that entails. All selection processes involve reducing uncertainty about the person being interviewed, who is usually subjected to all kinds of tests to assess their aptitudes, attitudes, character traits, etc. Similarly, employers have been looking at job candidates’ social network profiles for many years, working on the basis that if somebody behaves like an idiot on Facebook, they probably are an idiot. But using algorithms to do so takes things to a whole new level, and comes with the risk of including biases such as their inability to recognize or understand humor, irony or satire, or worse still, by interpreting them as dangerous. In the case of Predictim, when the algorithm identifies potentially dangerous attitudes, it marks those posts so they can be manually examined.

What happens when an algorithm of this type is used to decide who makes a good babysitter, who most of the time are young people? It analyzes their comments and behavior on social networks, which by definition are likely to be immature. Furthermore, if job candidates refuse to allow their social network profiles to be analyzed, their prospective employer is informed. An exhaustive analysis of somebody’s activities on social networks can be affected by other factors such as religious beliefs, political affinities, or even musical tastes. In the case of many managerial or technical positions, the issue is even more complex: many potential candidates participate in social networks to raise their profile in their area of ​​knowledge; from now on they may need to be more careful.

In case anybody out there had any doubts, algorithms of this type once again highlight the reality that what you do online stays online and may define you for life. If you spend your time on social networks pouring your heart out or pouring scorn on others, you are leaving behind a record of who you are for all the world to see. Needless to say, employers will see it and form an opinion of you during their selection process and may well decide to not to pursue the interview process further.

How many of us are perfectly happy for a future employer to read through all those angry tweets, those sarcastic comments on YouTube or our participation in the public lynching of someone? Oh, so you thought only politicians had skeletons in their digital cupboards?

(En español, aquí)