Talent Select AI automatically screens job candidates’ psych traits

If you’ve ever had to take what appears to be a personality assessment test when filling out a job application for a prospective employer, you’re not alone: “100 million workers worldwide take psychometric tests…designed to study personality and aptitude,” making for a $2 billion annual market, according to a recent article in The New York Times. Such tests are used by 80% of Fortune 500 companies to evaluate job candidates, according to Psychology Today.

The Myers-Briggs Personality test, one of the most widely used and recognized, contains prompts for self-rated responses such as “You regularly make new friends” and “Seeing other people cry can easily make you feel like you want to cry too” and takers are asked if they “agree” or “disagree,” and to what degree.

Talent Select AI, a 16-year-old digital interview and psychometric assessment firm headquartered in Milwaukee, Wisconsin, is aiming to disrupt the industry entirely with its natural language processing (NLP)-powered candidate screening tool, which does away entirely with self-reporting examinations from job candidates. Talent Select’s tool analyzes a prospective job candidate’s word choices alone during a live interview with a recruiter to conduct a psychometric assessment — using software to determine whether they are a good personality fit for the job opening.

A psychometric AI API

The company currently offers its software as an API to clients who can integrate it with their hiring platforms and tools. However, the company shared with VentureBeat that it plans to launch its own user-facing version of the software on its website next month.

“We’re new to offering this solution,” Talent Select AI CTO Will Rose said in a Zoom interview with VentureBeat. He said the company believes it will bring benefits to employers and job-seekers by eliminating the need for psychometric tests as a separate step of the interview process.

Rose said Talent Select’s AI model matches a job seeker candidate’s word choice and context of their conversation — using only a text transcript alone, no audio or video — to analyze their psychometrics and personality traits.

That’s because, as Rose points out, other types of tools that look specifically at visual information or intonation can have risks of bias against certain racial or ethnic groups. “We’ve seen a lot of pitfalls there because of the differences in cultures and how they are interpreted by others,” Rose noted. “In our case, we are looking strictly at the words.”

“We can determine just from the words how a candidate matches with a specific role or company culture,” Rose continued. “These are predictive in terms of job performance and job outcome.”

Promising initial results

Talent Select’s API launched earlier in 2023 promising “unbiased candidate insights” and has since resulted in initial improvements of a more than 50% reduction in time-to-hire candidates, an 80% increase in candidates selected from underrepresented groups and 98% of users report greater confidence in selection decisions.

Rose declined to specify to VentureBeat which clients Talent Select AI has so far serviced, citing confidentiality agreements, but said “we are working with existing providers.”

Psychometrics has a long history…but also criticisms and controversy

Psychometrics, the science of psychological measurement, has undergone significant transformation since it was founded at a laboratory at the University of Cambridge in 1887.

As the science evolved, the early 20th century saw psychometrics playing a key role in creating intelligence tests like the Stanford-Binet and Army Alpha and Beta tests for educational and military purposes.

Fast forward to today, and psychometrics employs advanced computer algorithms and complex mathematical models, such as item response theory (IRT) and structural equation modeling (SEM), to devise and evaluate psychological tests.

However, the field has been subjected to controversy and criticism over the validity and reliability of psychological tests across diverse populations and contexts, ethical and social implications of these tests for high-stakes decisions, and philosophical and epistemological assumptions underpinning psychometric models and methods. Even the creators of the famed Myers-Briggs test say that it shouldn’t be used to make hiring decisions.

Nonetheless, there remains a market for other comparable tools as evidenced above, and Talent Select AI believes it has developed a new, improved, more streamlined and efficient version.

The company claims to have 30 years of academic research and more than 15 years of in-house expertise in recruiting and hiring operations. Its leadership team includes president and chairman Stuart Olsten and COO Heather Thomas.

The company also maintains an advisory board of academics and psychometrics practitioners such as Drs. Michael and Emily Campion, Dr. Sarah Seraj and assistant professor John Fields.