MultiModel Redesign

 For my project, I want to focus on technological benevolence in the job market. I think it's a very important topic especially for those of us who are working toward a degree to help us secure a position and have a good career. I want to know how it will affect my chance of getting a position and what it might entail. For my unit project, I kept it straight and to the point but I think more information on the implications of technological benevolence in the job market would help the project stand more on its own.

In Race After Technology by Ruha Benjamin hits on the topic of using AI in the hiring process of a company she specifically talks about one company and how they implement AI during the interview process. She says, "A company called HireVue aims to "reduce unconscious bias and promote diversity" in the workplace by using an AI-powered program that analyzes recorded interviews of prospective employees. It uses thousands of data points, including verbal and nonverbal cues such as facial expression, posture, and vocal tone, and compares job seekers' scores to those of existing top-performing employees in order to decide whom to flag as a desirable hire and whom to reject"(Benjamin 96).

There are many companies now that claim to use an ethically designed AI to help analyze a person's interview and all the aspects of it to help sort through many applicants to decide who would be the best fit. Another one I came across in my research is called Pymetrics whose mission statement is “Our vision is to realize everyone’s true potential. Our mission is to make it happen with accurate soft skills matching.”

In the summer of 2020, Pymetrics did an audit to check their AI and its ability to be unbiased and it passed its audit. The University of Massachusets Amhearst did research on a couple of these companies and their results show that AI design can be unbiased and can improve the hiring process. To help ensure these programs are produced without bias Pymetrics enforces certain practices during the design process “After a data scientist has finished training, evaluating, and packaging a model for a given client they must complete a checklist that includes over 100 questions. These questions ask the data scientist to review all key aspects of the model building process, copy salient numerical data (e.g., accuracy and pass rates) into the sheet, and document in writing any significant deviations from the standard “templated” model training process. While much of this process could be automated, Pymetrics deliberately adopted a manual process that forces the data scientist to document and justify their work” (Wilson et al. 5.2). There is also a concern about privacy for the applicants, Dave Zelinski of SHRM says “as the use of AI has grown, it has attracted more attention from regulators and lawmakers concerned about fairness and ethical issues tied to the technology. Chief among those concerns is a lack of transparency in the way that many AI vendors’ tools work”. Pymetrics puts a focus on being transparent with all parties so as to provide applicants with the respect they deserve.

AI development is a process that is still being perfected and improved all the time. Using these programs to help sort through applicant pools to help select the best fit dies have good and bad implications first they can help analyze aspects of the process that we as humans wouldn't be able to. They could analyze speech patterns and facial cues and body language that a human could not pick up on. They would be able to do it without the first snap judgment as a human would. On the flip side, a big concern would be privacy infringement. How much personal data is part of this process and how can we protect a person's personal data in the process. The main company I research was Pymetrics who prided itself on ethical design and transparency. This leads me to believe that implementing AI in this process could be more beneficial than harmful. They claim to be completely open about what they are doing how they are doing and what it would be used for. I firmly believe that with the correct procedures in place and checks and audits along the way it is possible to design an AI that could be unbiased and neutral.

As a designer, I know that it is my responsibility to work towards more inclusive and unbiased design. Knowing the issues and understanding what's going on in the world and continuing education on world issues is the number one step in being able to design products that are unbiased and inclusive. Moving forward if we continue training in these practices even after formal education as well as put procedure and audits in place we can have a future where design could be more bias-free and allow for more people to have success.

 

Works Cited

 

Building and Auditing Fair Algorithms: a ... - Github Pages. https://evijit.github.io/docs/pymetrics_audit_FAccT.pdf.

Caprino, Kathy. “How AI Can Remove Bias from the Hiring Process and Promote Diversity and Inclusion.” Forbes, Forbes Magazine, 7 Jan. 2021, https://www.forbes.com/sites/kathycaprino/2021/01/07/how-ai-can-remove-bias-from-the-hiring-process-and-promote-diversity-and-inclusion/?sh=28e4c8cb4ec5.

“Center for Employment Equity.” UMass Amherst, https://www.umass.edu/employmentequity/using-technology-increase-fairness-hiring#overlay-context=making-discrimination-and-harassment-complaint-systems-better.

Heilweil, Rebecca. “Artificial Intelligence Will Help Determine If You Get Your Next Job.” Vox, Vox, 12 Dec. 2019, https://www.vox.com/recode/2019/12/12/20993665/artificial-intelligence-ai-job-screen.

“The New HR: How Ai Evolved the Hiring Process: Virtasant.” RSS, https://www.virtasant.com/blog/the-new-hr-how-ai-evolved-the-hiring-process.

          Princetonuniversity, director. YouTube, YouTube, 15 May 2020, https://www.youtube.com/watch?v=rY8RkET3KC0. Accessed 18 Dec. 2021.

“Pymetrics' Mission: Aligning Soft Skills to Careers.” Pymetrics' Mission: Aligning Soft Skills to Careers, https://www.pymetrics.ai/mission.

Zielinski, Dave. “Addressing Artificial Intelligence-Based Hiring Concerns.” SHRM, SHRM, 22 May 2020, https://www.shrm.org/hr-today/news/hr-magazine/summer2020/pages/artificial-intelligence-based-hiring-concerns.aspx.