You are a highly educated woman in your early thirties, you have years of relevant work experience, and excellent references. You lose your job due to a financial reorganization. How hard can it be to get a new job?
The answer to that question strongly depends on your last name, and your color. If your name is Smit, De Vries, or Jansen, there is a good chance that you can choose from job offers within a few weeks. If your name is Mohamed, Ahmed, or Çakır, the answer will be different. Very different. This is not just a ‘hunch’, it’s a proven fact. Between 2016 and 2018, researchers from UvA and Utrecht University sent out 4200 applications to vacancies that actually existed. They used fictitious, non-Dutch names. Results showed that if you have a white, Dutch background, your chances of getting invited to an interview is a lot bigger. To give a mindblowing example: people with a Turkish, Maroccan of Antillean background have 40% smaller chance of getting a response.
We hear these stories first hand now. One of our new colleagues, the woman from the beginning of our story, told us that she spent months and months looking for a job. Day in, day out. The chance that she got a reply to a cover letter was less than 10%. Even though she met all the criteria, the probability that she was invited was less than 5%.
What is the solution? Do we need to implement a quota, just like we did for women? Is anonymous job application the way out? The fact that no one will know your age, background or name, seems to be an obvious advantage. The thing is, however, it becomes a disadvantage if you want to work on diversity and give people who are deprived a better chance. And what will happen if the applicant shows up at an interview? Will there not be delayed discrimination in a lot of cases? And, who exactly will you be inviting if applicants are forced to leave a large part of their identity behind?
Is AI the solution? More and more recruitment and selection tasks are automated. Take for instance Vera. Vera is a recruitment robot. According to her manufacturer, she can select candidates 10 times faster than her human counterpart. She saves up to 60% of the time a recruiter would spend on the hunt for talent. And to go back to discrimination: Vera (or AI in general) does not have bias, because it’s a system. The question is, however, if that is completely true. AI works with information that is selected and put in by humans. Therefore it more often than not contains a lot of bias, whether that is consciously or unconsciously. And Vera, of any other system, will take over that bias. Which eventually still leaves you with people with smaller chances; they only get excluded 10 times faster.
Awareness and action
Essential in this whole story is awareness. Admitting there is discrimination. Recognizing it and taking action against it, for starters by checking you data for bias. Because, isn’t it insane that people do not get a chance to show their talents, just because they do not have your typically Dutch name and are therefore met with a whole series of absurd prejudices? And they are not only absurd, they are dangerous too. Think about what would happen if Dutch politicians would make crazy statements that are full of preconceptions and that would be adopted by others? Just imagine...
Oh, wait, how about the apps in which one of those politicians states that on average white people score 110 on the IQ scale, Latinos 90 and Afro-Americans 75? One of those politicians who could be in control of the future of our country? The representatives of the people? In March, there will be elections. That will be the ideal moment to deny these characters any speaking time or influence whatsoever. Shall we?
On February 11, we will place a pagelong ad in FD about this subject. You can find it below.