Pages

More details on how AI is responsible for shortlisting names and how LinkedIn took care of its biased AI program

There was a saying that Artificial Intelligence is getting so advanced day by day that in near future it will take over the jobs that were previously done by real people. Now it can be said that AI will now decide whether someone is fit for a job or not.

After it was reported that the AI program used by LinkedIn is biased and promotes male profiles over females, the AI program was then replaced with a new one. While LinkedIn was dealing with its AI program, other such job finding websites were not open on how exactly their AI program works, on what grounds it decides whether someone is suitable for the post or not. These sites include Career Builder, Monster and Zip Recruiter etc.

AI is getting embedded with the recruiting process, many companies now make their decisions based on the results of the AI games they ask their employees to play. Similarly it will decide whether a profile be made available to the recruiter, this decision is based on the type of searches done by the profile. Use of artificial intelligence became more significant when Covid19 struck the world. Since one on one interviews stopped for a while, new recruitments were done online. It was reported that companies who offered AI powered interviews observed a rise in their earnings.

These websites are playing an important role to help people find a job they are qualified for and Artificial Intelligence is the reason why it is happening. According to the Vice President of Monster, Derek Kan, there is a phrase that only 6 seconds are given by recruiter to review a CV. When it comes to AI program, milliseconds are enough to shortlist the names.

The VP of LinkedIn, John Jersin told that AI shortlists the candidate based on three cases. The data fed by the job seeker. The data given to be used in comparison to others with similar skills, interests of the user along with their pattern of using the site. The AI program of LinkedIn doesn’t ask for name, age, gender, age as it is feared that results might get biased. But the AI program is too smart and can easily detect behavioral pattern.

As it is believed that men are more serious while looking for a post to work on even if it is above their qualification, while for women, they only opt for posts they are qualified for. This gets noticed by AI program based on users behavior and as a result women face disadvantages. Such as man getting notified with jobs more than there qualification whereas woman are only restricted to jobs according to their qualification. These were the issues that were pointed out by LinkedIn and lead to build a new AI program which was launched in 2018. The new program now unsure an evenly matched distribution of available jobs across gender.

While talking about LinkedIn, Derek Kan told that unlike LinkedIn who fixed their biased result through new AI program, Monster doesn’t do anything to prevent such results on their platform instead they attract users to get linked with them and take advantage of the services.

The Chief Executive Officer and cofounder of Zip Recruiter, Ian Siegel told that just like LinkedIn, they do not take names under consideration instead, they rank them based on 60+ different types of categories which even includes their address with respect to geographical location. Ian Siegel was confident that his company ranks the job seekers more efficiently than others and all of this is done on the basis of merit.

It is a tough challenge for job seekers to handle both AI programs and recruiters to prove themselves to be worthy of the position. The challenge gets more tricky as AI will keep getting more involved in this process. This is why it is important that how the job finders are showcasing themselves on such sites. The better they do this, the more chances they will have to get selected by the system’s algorithm.

H/T: MIT TechnologyReview.

Read next: Google Attempts to Address Racial Disparities With New Skin Tone Scale

No comments:

Post a Comment