We’ve all been through the nerve-wracking position of interviewing for a job. At HireVue, we know from both data and personal experience that the old process of finding, vetting, and hiring candidates is broken, and it’s broken in large part because of the systematic and unchecked influence of both conscious and unconscious human bias throughout the process. Here are 4 required steps for creating tools for equitable hiring solutions.
1. Acknowledge the Pervasive Discrimination in Legacy Hiring Practices — Understanding the Problem Is the First Step Toward a Solution
Conscious and unconscious discrimination and bias in hiring is pervasive, a fact that is borne out in recent research in both the United Kingdom and the United States. Since our founding in 2004, HireVue has been focused on democratizing the hiring process globally and making it fairer and more accessible for all candidates. We’ve built a team of people who are driven to fight the influence of human bias and democratize the hiring process.
HireVue’s first solution to mitigate bias was the structured OnDemand video interview, which opened up the opportunity for all interested candidates to showcase their skills with a video interview. The ability to deliver a consistently fair interview for all candidates in the process is the first important step in improving the process for all involved. One of our customers, a children’s hospital in the Midwest, increased the diversity of its hires 55% by offering these video interviews, so candidates could “tell their stories” and be more than words on paper.
2. Require Evidence-Based Validation for All Good Ideas — How Does It Fare in the Research?
It’s not enough to simply have a “good idea” and then automate it at scale: our team of scientists draw on a long, well-established history of predictive assessment science for employability that has been tested and refined for more than 100 years. The people behind these assessments are industrial-organizational (IO) psychologists, academics who specialize in the study of human behavior in organizations and the workplace. One of the key roles of an IO psychologist is to help develop criteria and assessment tests to evaluate performance, and since the establishment of the 1978 Uniform Guidelines by the Equal Employment Opportunity Commission, to ensure that those performance criteria are job-related and not perpetuating bias.
Our video and Game-Based Assessments were introduced a full year after our team of IO psychologists pioneered a way to perform bias testing that meets or exceeds the requirements of the Equal Employment Opportunity Commission’s Uniform Guidelines (often called the “4/5ths Rule”). This is considered the gold standard for the prevention of bias in pre-hire assessments (those multiple-choice tests you may have taken in the past) and it is the testing guideline for all of our assessment models before anyone of them can ever be put into production.
3. Third-Party Input Is a Must — Technology Creators Have a Responsibility to Look Outside Their Conference Rooms for Expertise
We’ve gone outside of our organization to augment expertise. For instance, our Expert Advisory Board directs our product team on key aspects of HireVue technology development, including AI ethics, algorithmic fairness, diversity and inclusion, and data security and privacy. It is also why we regularly invite data science researchers, social scientists, educators, and others to interactive roundtables, where we present extensive detail on our development and bias-mitigation process and invite informed dialogue.
We are also working with Integrate Autism Employment Advisors, who are using HireVue Assessments to evaluate and coach more than 600 candidates in their network who are on the autism spectrum to be placed with companies in the Fortune 500 and beyond. This partnership is critical for people with autism who experience rampant employment discrimination. Our partnerships with groups like re:work and Integrate Advisors are just some of the ways we’re working to ensure our technology is helping build a more inclusive job market where qualified candidates with a diverse set of competencies, abilities, and disabilities receive the opportunities they deserve.
4. Artificial Intelligence Is a Tool to Help, Not Eliminate, Human Involvement — None of Us Want Unchecked Automation
In the traditional process for large organizations, recruiting teams are so inundated with resumes that recruiters are only able to give a cursory glance to resumes that applicants spend significant time, and sometimes money, to create. Recruiters consciously scan for keywords, recognizable school names, and a list of skills to see who moves on to the next step in the interview process. Unfortunately, interviews — as humans — also often unconsciously eliminate names that don’t sound white and male.
It is this bias that our tools can help address, but our goal has never been to eliminate human participation; it’s to focus humans in the recruiting and hiring process on the things at which people excel, such as relationship-building and negotiation. Once the most diverse set possible of candidates with the right competencies for the job in question has made it past the screening phase, skilled hiring managers and team members can now have one-on-one conversations with candidates who might otherwise have been put into the proverbial trash bin for reasons that never should’ve been considered in the first place.