Humane Algorithms - can they happen?
If 5 HR people work 12 hours a day (including weekends) and spend five minutes on each application, how many days will it take to go through quarter of a million resumes? About a year, is the answer. That’s the number of applicants Goldman Sachs got two years back. They have no choice but to rely on an Applicant Tracking System (ATS) algorithm that looks for key words and finds a match. Candidates must get past the algorithms in more and more companies.Software systems can in some cases be so efficient at screening resumes and evaluating personality tests that 72% of resumes are weeded out before a human ever sees them.So what do candidates do to beat hiring algorithms? They use multiple resumes with multiple keywords. Some innovative ones put keywords in white on the resume that are invisible to the human eye but are read by machines. Machines are great at following rules. Humans are ingenious when it comes to breaking rules.
Algorithms are codified biases
Algorithms speed up our decisions when the choices are not very different from each other. As a report from Pew Internet says rightly, “(algorithms) put too much control in the hands of corporations and governments, perpetuate bias, create filter bubbles, cut choices, creativity and serendipity, and could result in greater unemployment.” After all they are only opinions expressed in code. They carry all the biases of the ones coding and the bias of the data sets that they draw upon.
Gender is not the only bias
The Economistspeaks of the challenges of ads promoting jobs in science, technology, engineering and math on Facebook. “They found that the ads were less likely to be shown to women than to men.This was not due to a conscious bias on the part of the Facebook algorithm. Rather, young women are a more valuable demographic group on Facebook (because they control a high share of household spending) and thus ads targeting them are more expensive. The algorithms naturally targeted pages where the return on investment is highest: for men, not women.”Knowing how algorithms work is an essential part of getting hired. Given the scale, importance and secrecy of hiring algorithms, they have the potential to create a category of people who will get rejected and never know why. There are algorithms that find out the weather forecast and only then decide on the work schedule of thousands of people who do these gigs.
Algo algo everywhere
Search engines, the apps on your phone, dating sites, job sites, news sites, shopping sites, travel sites all run on algorithms. Computer and video games are stories told through algorithms.Every time that you click a button, the action becomes a data point that an algorithm could use to make your life more efficient. 125m households watch Netflix for more than two hours a day on average. Each pause, rewind, downloaded but unwatched movie is a data point that is used to classify the viewer into one of the 2000 “taste clusters” that Netflix uses. They encourage every family member to create their own profile so that even within a family, the minor variances in viewing habits can be used to make their algorithm more efficient in making recommendations that are harder to resist.
Surveillance capitalism
Shoshana Zuboff calls it “surveillance capitalism” where companies know so much about you that they can nudge you to buy more when you can least afford it. Or binge watch a serial and get sleep deprived just before an exam. Sarah Kessler, writes in Gigged how Uber employs hundreds of social scientists and data scientists to help manage the drivers through their app. They use videogame techniques, graphics and non-cash rewards to nudge drivers to work more hours.
Efficient but not humane
One of the budget airlines I traveled in asks all passengers to declare in advance how many kilograms their bags will weigh. At the terminal, if the bag exceeds the estimated weight by a kg, the excess baggage penalty is severe. Watching an old lady in a wheelchair plead with the airline to waive off the penalty is heartbreaking. It is surely efficient, but certainly not humane.Take the case of Xerox. Algorithms found that job applicants with long commutes are more likely to churn. But Xerox managers noticed another correlation. Many of the people suffering those long commutes were coming from poor neighbourhoods. So Xerox, to its credit, removed that highly correlated churn data from its algorithm. The company sacrificed a bit of efficiency for humaneness.Just as we care if our food is organic or if the shoe has been made using child labor, we must start asking how the algorithms are being supervised.
The value of getting lost
Algorithms make us a prisoner of the past. The algorithm serves you a filtered set of choices based on your demographic profile. Once you make a choice, it has more data points to limit the options that you can choose from. It stops us from experimenting. We start choosing from a narrow set of options and stop trying to learn from our failures. Relying on algorithms lowers the general population’s ability to make decisions.Algorithms assume that we will follow the patterns we have created through our decisions. Humans have to use their creativity to try what they never have. Curiosity is the antidote to bypassing algorithms. The more unpredictable your choices are, the less machines will control you.If we are not careful, we will move towards a society where one group will write algorithms and the rest will be ruled by them. That is the biggest reason to periodically go out of the defined path and get lost. Only then will we stay human.
====Written for the August 2018 issue of People Matters