Has Automation Gone Too Far with Facial Recognition?

Automation tools designed to reduce or eliminate high-volume tasks are swiftly transforming the workplace. And while this technology shift has been more pronounced in some industries over others, it’s particularly taking hold in the human resources (HR) space. 

Reviewing resumes and pre-screening applicants to identify well-suited candidates can be a draining and meticulous tactic, and it is particularly difficult to scale these efforts without the aid of automation tools. Therefore, businesses continue to slowly adopt hyper-intelligent software, including artificial intelligence (AI) and machine learning (ML). 

From smarter worker scheduling and talent management to employee onboarding, technology such as AI and ML makes it easier for companies to efficiently recruit, support, and retain workers while keeping costs in check. 

The adoption of AI by businesses increased by 270% in just four years and is projected to reach $266.92 billion by 2027. Interestingly enough, the human resources field is contributing to this growth.

In a 2021 Virasant study, 52% of HR leaders said the most challenging part of the recruitment process is identifying the right candidates from a large applicant pool

AI Strengthens the Hiring Process

AI in the hiring process helps HRs with time-consuming tasks, from mundane administrative tasks to improving job matching processes and speeding up the time it takes to screen, hire, and onboard new candidates.

And most notably, AI-powered facial analysis software can interpret desirable indicators from recorded video interviews and supports hiring managers when choosing high-quality talent.

Facial recognition software represents a significant positive shift in technology today. In summary, facial recognition is a biometric technique that uniquely identifies a person’s face.

For many, facial recognition isn’t anything new. Many of us unlock our mobile phones through the facial recognition features installed within our Apple devices without even thinking twice about it.

But in HR, facial recognition technology can provide a clear overview of an employee's working hours without having to add people's shifts and clock-in times manually. And this technology is also useful for managing the flow of people entering and leaving offices or warehouses. 

Challenges of Analysis Technology in the Hiring Process

Pushing against these trends, however, are concerns around AI in hiring. Specifically, experts have voiced legitimate concerns about facial recognition and analysis technology.

The use of facial analysis technology to evaluate job candidates is "very problematic," Frida Polli, founder and CEO of the New York-based assessment company Pymetrics, said in a recent press release. "The science of the technology in terms of what it really says about someone is extremely new and not well-validated, and certainly not well-validated for HR uses.”

Employers utilizing AI to make hiring decisions based on facial appearance, expressions, or emotions are also very possibly biased and unreliable. And due to the uncertain landscape, some favor pumping the brakes on facial recognition software altogether. 

In a letter to Congress, IBM CEO Arvind Krishna wrote, “IBM firmly opposes and will not condone uses of any technology, including facial recognition technology offered by other vendors, for mass surveillance, racial profiling, violations of basic human rights and freedoms or any purpose which is not consistent with our values and Principles of Trust and Transparency.”

And there is real substance to racial concern as well. Face recognition algorithms boast a high 90% classification accuracy,  but these outcomes are not universal. A growing body of research exposes error rates across demographic groups, with the poorest accuracy consistently found in subjects who are female, Black, and 18 to 30 years old. 

E-commerce giant Amazon also provided a contemporary lesson in what can go wrong when incorporating AI tools in the hiring process. In 2018, BBC News reported that Amazon’s automated system taught itself that male candidates were preferable and penalized resumes that were more likely to belong to women. The company has since abandoned the practice

Avoiding Unintended Bias in Facial Recognition 

The simplest way to avoid unintended bias and other possible ethical issues around facial recognition would be to take the facial analysis completely out of the hiring process. 

And this is just the approach that Wisconsin-based Talent Select AI has taken with its interview analysis platform. Talent Select’s platform provides hiring managers and recruiters with outcome predictions and insights about job candidates. 

Outcome predictions include hiring decisions and future job performance, while the platform’s insights include, but are not limited to, personality traits, emotional traits, behavioral traits, performance skills, competencies, and other non-cognitive traits. 

Notably, the platform avoids issues around facial recognition and voice analysis by only analyzing the transcribed text of interviews. In other words, Talent Select AI is a text analysis company, not a video analysis company.

Instead of applying facial analysis to recorded video, the company transcribes the audio portion of the recordings and leverages machine learning to identify success indicators. 

“Employers are worried about bias, and we agree that the science isn’t there to accurately analyze faces or inflections in speech, which can vary by culture and context,” said William Rose, chief technology officer at Talent Select AI, said in a recent interview. “Our algorithms and intelligent software are concerned with the content of answers, not how they are said. The video content, on the other hand, is still a valuable tool for real-life human recruiters to assess factors such as presentation and soft skills, which are important for jobs that require interaction with customers, for example.” 

As with any move toward automation, there must remain a human element somewhere in the chain. The problem with AI in the hiring process is that we may need to rethink where that line should be.