The Online Job Search — the Concerns and What You Need to Know

In our digital world, many people choose to utilize online job search platforms like LinkedIn and Indeed. While many users hail these platforms as doing social good by connecting people with opportunities, are they helpful to everyone? Job search algorithms block opportunities for certain qualified job seekers because they model gender and racial biases that are present in our society.

Gender bias has been a problem with job search algorithms because gender bias is a large societal problem. Image from Sheelah Kolhatkar with the New Yorker. URL: https://www.newyorker.com/magazine/2017/11/20/the-tech-industrys-gender-discrimination-problem

Luciano Floridi argues in “Ethics After the Information Revolution” that information technologies and digital technologies reduce friction that happens in information exchange or distribution. Meaning that as technology advances, there will be fewer barriers to information. However, with the growth of hiring algorithms, certain groups of job seekers are not being shown all the opportunities they are qualified for. Is technology that blocks opportunities for qualified candidates frictionless? As I see it, this environment is not frictionless. As Miranda Bogen explains, an extremely effective barrier to people finding jobs is not notifying them of opportunities. If technology reduced the friction in information exchange, then everyone would be shown all the job postings they are qualified for. Examining these points, the argument that biased job search algorithms contribute to the creation of more friction in information exchange is confirmed.

Image from Mike Egan at Insider Pro. URL: https://www.idginsiderpro.com/article/3544019/zoom-fatigue-is-real-and-it-s-costly.html

While job search algorithms appear helpful to most, they can further engrain certain biases within society and the hiring process. Algorithms learn from employers’ actions and preferences, then send out targeted ads to “relevant” job seekers (Bogen). A study conducted by Harvard Business Review found that “most hiring algorithms will drift toward bias by default”. Further research highlights that algorithms can “mimic subconscious human gender bias without taking into account the merits of a candidate”. This begs the question, who is at fault and how do we fix it? The humans that make the algorithms are not solely to blame, as societal structures broadcast norms that algorithms utilize to disseminate opportunities. Using large and diverse training sets for algorithms might seem like a solution, but the real world is biased, so this would still create a biased algorithm.

What should be done to correct these algorithms? Is it even possible to reduce bias in algorithms? There are many questions to ask, and not many people to answer them. I have only scratched the surface in this post, as this is a very complex problem. Improving algorithms will not completely resolve societal problems, but it would reduce friction and be a positive start.

References:

Bogen, M. (2019, May 6). All the Ways Hiring Algorithms Can Introduce Bias. Harvard Business Review. https://hbr.org/2019/05/all-the-ways-hiring-algorithms-can-introduce-bias

Floridi, Luciano. (2010). Ethics after the Information Revolution. The Cambridge Handbook of Information and Computer Ethics (pp. 3–19). Cambridge University Press.

Krishna, T. H. (2020, December 2). Entry barriers for women are amplified by AI in recruitment algorithms, study finds. Phys.org. https://phys.org/news/2020-12-entry-barriers-women-amplified-ai.html