How biased AI systems omit hidden workers and cause a talent shortage

It should come as no surprise that one of the most influential technologies of the 21^st century, artificial intelligence (AI), is also widely by recruiters as a tool to attract, screen, and engage candidates for open positions.

But what if human bias is affecting how we create algorithms for AI systems?

Suggested reading on hiring bias: Combatting Bias in Machine Learning Algorithms: For Recruiting

Here’s how biased AI systems are causing a talent shortage leaving behind a large pool of hidden workers.

AI and the hiring process

Recruiters rely on AI technology to assist with the hiring process in the following ways:

1. Speed up the time it takes to manually screen each candidate
2. Eliminate human bias or assumptions that prevent the best candidate from being selected for a role

For recruiters, AI excels in its time-saving capabilities in the hiring process. Algorithms and machine learning make it quick and easy to screen resumes, automated assessments, and schedule interviews. This all empowers hiring managers to fill open positions faster.

However, a closer look reveals that unless your company is careful with its AI-enabled hiring process, algorithms can do more harm than good when it comes to your commitment to diversity, equity, and inclusion.

What is AI hiring bias?

While advocates have boasted that AI technology will be a key solution to ending human bias and ruling out discrimination in the hiring process, not all recruiting algorithms are created equal.

Unless algorithms are trained with data representative of candidates from all demographics (and in equal proportions), flaws in the algorithms have the potential to create biases in every step of the hiring process, from talent attraction to screening and beyond.

Here are a couple of ways biased AI systems can create diversity barriers and narrow a candidate pool:

Talent attraction

Insourcing talent for a role, recruiters use the predictive technology of AI to advertise open jobs to the right talent and identify qualified applicants. By using these algorithms to target candidates, predictions are made based on which applicants are most likely to click on the ad, rather than which candidates would be the best fit for the job.

The issue with these predictions lies in how these job ads are delivered to potential candidates. In creating targeted ads on a social platform such as Facebook, biased AI systems can reinforce gender and/or racial stereotypes, among other biases.

Screening resumes

One of the most popular uses of AI technology in the hiring process is the automation of screening resumes with algorithms.

AI technology is trained by humans and relies on past hiring data to identify features of a good or bad candidate. If prior biases existed within your organization, you’ll run the risk of the system adopting specific attributes that are only common among your current workforce.

As an example, if most of your organization’s employees are males, the algorithm may learn from the existing data set to prefer male candidates over female, non-binary, etc. When an algorithm learns to filter candidates based on characteristics or identity traits rather than job competency, there’s a danger of exhibiting unintentional biases.

How do biases affect recruiting?

Conscious or unconscious biases that exist in the hiring process can lead to discrimination, preventing you from creating an inclusive and diverse work environment. Biases can cause your organization to make bad hiring decisions, preventing you from finding the best candidates to fill open positions.

Biased AI systems recruiting also have harmful side effects that can ripple far past your organization’s diversity initiatives. In addition to higher turnover and lower retention rates, bias can negatively impact employee morale and engagement.

The bottom line is biases limit a talent pool both in size and diversity. And in today’s hot labor market, narrowing your candidate pool with biased recruiting can be costly.

The talent shortage caused by biased AI systems

While there’s no denying the competitive landscape of the labor market, AI recruitment can further shrink your candidate pool when used to advertise job postings or sort resumes.

In fact, most algorithms are trained to separate candidates that are assumed to be “unfit for hire,” rather than pinpoint the applicants who would be most successful in the role. The result is an overlooked pool of talent, also referred to as “hidden workers.”

Hidden workers are unemployed or underemployed individuals who are eager to work. They may include:

• People who are underemployed and working part-time jobs while willing and able to work full-time positions
• Applicants actively looking for work but who have recently experienced extended unemployment caused by several reasons such as health issues or family care responsibilities
• Full-time workers who are not actively looking for a new position, but may consider a change given the right circumstances

Tapping into this overlooked category can lead to great benefits for employers. Along with fostering diversity, equity, and inclusion, companies that hire hidden workers are 36% less likely to face skills and talent shortages compared to organizations that don’t.

The backgrounds of hidden workers

The authors of Hidden Workers: Untapped Talent found that hidden workers come from very diverse backgrounds, including:

• Caregivers
• Veterans
• Immigrants and refugees
• The physically disabled
• Partners of relocated workers
• People with mental health or neurodiversity challenges
• People from less-advantaged populations
• Previously incarcerated individuals
• People without traditional qualifications

Suggested resource on talent hunt: How you can uncover opportunities in the new talent landscape

7 strategies to help you find the right talent for your role

Today’s labor market is competitive, but filling positions with strong applicants is by no means out of the question. You need to focus on overcoming the biggest recruitment problems to find the right talent for your role and build a diverse workforce.

Here are seven strategies to improve your recruitment process and secure talent in even the toughest labor market:

1. Rethink experience requirements and look for potential applicants that possess transferable skills and raw potential.
2. Improve candidate engagement with automatic and personalized messaging.
3. Speed up your hiring process to identify top candidates faster than your competition.
4. Pursue non-traditional candidates or hidden workers who are filtered out by AI.
5. Build a high-quality candidate pipeline with potential candidates who are ready to move when you are.
6. Improve your interview processes by creating a predetermined list of questions to ask all candidates.
7. Use an AI recruitment system like Arya with built-in diversity to create an inclusive workforce for your organization.

Ready to find those hidden workers?

Finding the right AI solution to combat biased recruiting is critical. Diversity programs are no longer optional for businesses of all sizes.

At Leoforce, we understand the need for data-driven hiring decisions and reducing bias in recruiting procedures. Our AI recruiting platform, Arya, keeps learning and improving, reducing human prejudice and boosting diversity initiatives.

Unlock the universe of talent and see Arya Quantum in action by requesting a personal demo.

Resources

• https://hbr.org/2020/12/how-businesses-can-find-hidden-workers
• https://www.accenture.com/_acnmedia/PDF-163/Accenture-Is-the-Talent-You-Need-Hiding-in-Plain-Sight-Final2.pdf