Human, AI Biases Haunt Recruitment Processes
The sudden implementation of AI on recruitment processes might have accelerated hiring but, if programmed incorrectly, AI could be a double edge sword. Mexico has limited comprehension of this tool, explains IBM, further increasing risks.
AI will play a leading role in the new working era because it is one of the trends that transformed the world during the COVID-19 pandemic, explains Selene Diez, CEO and Founder, Forte Innovation Consulting. “AI has demonstrated the capacity to transform interactions in all aspects of our lives. At least 50 percent of employees are immersed in everyday tools that capture their data.”
For recruiters, AI and machine learning enable recognition patterns, catalog behaviors and evaluate the employees’ reasons for leaving a company, which is one of the most common problems companies face, said Diez.
About 20 percent of Mexican companies embraced AI after the pandemic, but many of them rushed to integrate the tool into their operations without addressing why they implemented it, according to an IBM report. Out of that 20 percent, less than 5 percent of companies use data services for candidate management, said Pol Morral, Co-founder and CEO, LaPieza.io. “As a result, companies take an average of 40 days to hire and consider only experience, education institutes or acquaintances, instead of future potential or reskilling,” Morral said.
The benefits of AI can be overshadowed by poor algorithms that create biases throughout the recruiting processes. Any human bias that may already be in a recruiting process, even if it is unconscious, could be learned by an AI developed without due diligence. This happened to Amazon in 2015, according to Reuters, when the company realized its new system was not taking a gender-neutral approach to rating candidates for software developer jobs and other technical posts. Amazon’s AI was developed under a male-dominated the tech industry, so the system taught itself that male candidates were preferable.
However, if properly developed, AI can eliminate unconscious human bias, said the Harvard Business Review. AI should be audited so bias found can be removed. If standards are not met, the defective technology must be fixed before it is allowed into production.
To avoid these biases, expert Monica French, Head of New Business Hispanic America, LinkedIn Talent Solutions, urges proper regulations so companies can revise their processes to identify what can be limited, what should be regulated and what practices should be left behind.
“Most importantly, the final responsibility falls on the human element because technology is just a tool,” says French. While AI has imperfections, it is up to people to limit them. AI can reduce hiring mistakes that cost companies large figures, inhibit project advancement and sink progress, she added.