paint-brush
How AI’s Bias Problem Can Affect Your Next Job Applicationby@amply
10,871 reads
10,871 reads

How AI’s Bias Problem Can Affect Your Next Job Application

by AmplyMay 22nd, 2024
Read on Terminal Reader

Too Long; Didn't Read

AI bias in recruitment arises when systems and tools that are used to screen and select candidates, inadvertently perpetuate or even amplify existing prejudices. These biases often stem from the data these systems are trained on, which can reflect societal prejudices or historical hiring patterns. Names distinct to Black Americans had a lower likelihood of being listed as the best applicant. Names connected to Asian women and white men generally performed better.
featured image - How AI’s Bias Problem Can Affect Your Next Job Application
Amply HackerNoon profile picture

Tech job seekers are being blocked by AI bias when applying for roles. Though for HR professionals, gen AI has plenty of potential when it comes to automating tedious tasks, its tendency towards bias remains a cause for concern.


AI bias in recruitment arises when systems and tools that are used to screen and select candidates, inadvertently perpetuate or even amplify existing prejudices.


These biases often stem from the data these systems are trained on, which can reflect societal prejudices or historical hiring patterns that favor certain demographics over others.


Sometimes AI bias is due to the design itself, reflecting unconscious biases of its developers, while feedback loops mean that if a system consistently chooses candidates from similar backgrounds, it will continue to do so unless corrected.


Inbuilt bias can impact your job search significantly. While reporting has been widespread about biases on Applicant Tracking Systems (ATS), which scan resumes for keywords in the job spec and for keywords that align with previous successful candidates, a recent experiment shows just how bald AI prejudice can be.


Bias in action

By applying fictitious but “demographically-distinct” names to equally-qualified resumes, ChatGPT 3.5 was asked to rank the applicants against a real job spec for a financial analyst role with an actual company.


Names distinct to Black Americans had a lower likelihood of being listed as the best applicant; names connected to Asian women and white men generally performed better.


With the dizzying speed of AI development, some say this experiment is already dated. Unfortunately, that’s not the case.


Though ChatGPT-4o is now being rolled out, and OpenAI claims to have tested it with outside specialists, it also says it will keep addressing concerns like bias, fairness, and disinformation as they come to light. Meaning biases still exist.


What can job seekers do to circumvent AI bias?

For job seekers, a little digging into a company’s commitment to DE&I will indicate if it is likely to be focused on de-biasing its AI tools.


With the ability to scan thousands of resumes in seconds, it’s unlikely that companies will return to only manually reviewing resumes, but many organizations are adding a layer of human oversight in screening processes to help mitigate against biased AI results. Of course, humans have biases too, so no method is foolproof.


Other hiring teams are using ATS that hide applicants’ names and education levels.


Also, look to see if your desired company is hung up on having a particular education level; there’s a higher chance of AI bias if it is.


Of course, in particular industries like medicine and healthcare, professional qualifications are very important, but for others, like software engineering, organizations can take a more skills-based hiring approach.


Engineers and developers are often STARs (skilled through alternative routes), so look out for experienced-based requirements over educational minimums in job ads.


Ready to start your job hunt? Visit the HackerNoon Job Board, which contains thousands of interesting tech roles, like the three below.


DevOps Engineer, SAP, Bellevue

Sap is seeking a DevOps Engineer to be responsible for the effective and efficient delivery of customer environments in Cloud of respective product lines, and to ensure delivery is according to defined Service Level Agreements.


You’ll also contribute to the development of automated solutions for anomaly detection and incident auto-remediation, monitor and measure systems for performance, load and capacity planning, and ensure smooth operations and maximize uptime.


To be considered you’ll need at least two years’ experience in Docker, Kubernetes, Jenkins, Ansible, Python and BASH among others. Find out more here.


Senior Software Engineer - DevOps, Roku, San Jose, CA

Offering a sizable salary, though for California-only, Roku is hiring for a Senior Software Engineer with exceptional DevOps skills. Responsibilities include automation and scaling of Big Data and Analytics tech stacks on Cloud infrastructure, building CI/CD pipelines, setting up monitoring and alerting for production infrastructure, and keeping its tech stacks up to date.


With eligibility for health insurance, equity awards, life insurance, disability benefits, parental leave, wellness benefits, and paid time off, this role requires four years of experience in DevOps and/or reliability engineering.Read more here.


Senior Staff DevOps Engineer, FanDuel, Atlanta

Play a pivotal role in automating FanDuel Group’s cloud infrastructure and applications by collaborating to software engineers, release engineers and other cross-functional teams as a Senior Staff DevOps Engineer. The successful applicant will drive key initiatives by developing, planning, and implementing strategies across multiple cross-functional areas that align with its business objectives.


While the primary focus is on AWS, other cloud platforms may also be required, so showcasing a commitment to professional development in cloud tech is important. Progressively, no educational requirements are listed, but a list of desired experiences and skills are. See requirements here.



By Amanda Kavanagh