“Not only do we give them knowledge of which gender applies, [the candidate’s] With AI screening scores, there are no gender differences in evaluation,” Leibrand said.
At the same time, the study also found that informing candidates that their evaluations would be judged by computers instead of humans reduced the number of women completing applications compared to men because they believed women would be judged more fairly. was found to increase by about 30 percent. by AI rather than by humans.
Adoption of AI
Overall, the study found that using AI in recruitment almost doubled the number of women in the top 10% of performers. This means that the use of AI could lead to more women being recruited for similar types of technical jobs.
“To reduce gender diversity, AI tools would need to be heavily biased towards women.” [in the male-dominated technology industry] than it would have been discovered without AI,” the study concludes.
Dr. Mallorie Avery and Associate Professor Joseph Vetch and co-authors of the study, Professor Leibrand, said the study’s findings suggest that using AI as an early screening tool will help employers in other industries eliminate biases in the hiring process. said to suggest that it could help reduce
However, as this was the first study of its kind and focused solely on the tech industry, further research is needed to confirm the findings, he said.
The study, based on two experiments, is believed to be the first academic study to explore how the use of AI in recruitment directly impacts both supply and demand for minority job seekers. It is
In the first experiment, the researchers posted web designer job ads and invited more than 700 interested job seekers to apply. Some applicants were told that their applications would be evaluated by AI, while others were told that they would be evaluated by humans. The researchers then measured application completion rates to determine whether informing applicants that they would be screened by computer would attract women or discourage them from completing applications.
Then, two follow-up surveys of job seekers allowed researchers to understand why women were more likely to apply for jobs as evaluated by AI.
A second experiment focused on the behavior of human recruiters. The researchers enlisted more than 500 technical experts to act as recruiters for the web developer role, and found that “these professional evaluators have access to applicant evaluation scores provided by AI software.” and whether the applicant’s evaluation score can be inferred” was randomized.sex [from their names]”.
“This shows how supply and demand converge to create an overall change in the diversity of the pool of applicants who are most likely to get the job – those who sit at the top end of the ratings distribution. It will be possible to assess whether it occurs,” the researchers wrote.
“I have the ability to break down prejudices.”
Barb Hyman, Founder and CEO of Sapia.ai, said the independent study demonstrated the use of ethical AI in recruiting is “the most comprehensive recruiting method.” said.
“In 2022, women will only make up 27 percent of the workforce across the STEM sector, down from 2020,” Hyman said.
“This study shows that ethical AI systems can break down prejudices and remove many of the barriers women face when entering the tech industry.
“Recruiting on the basis of soft skills, behavioral traits, and cognitive abilities alone will result in hiring as many women as men, if not more. They are all irrelevant factors in determining whether
Sapia.ai uses AI-powered chat interviews to screen applicants and provide employers with shortlists of candidates. Other of his AI-assisted recruitment tools include HireVue, Humanly, HireScored and Paradox.ai.
Sapia.ai’s tool allows candidates to answer five questions in their own time, without having to reveal their gender, age or race.
The company’s software is trained on a database of 1 billion words of structured interview responses from 2.4 million candidates in 47 countries, and uses machine learning and natural language processing to interpret those responses. Read and assess the candidate’s suitability for the job. Each candidate is given her score out of 100.
Candidates will also receive an email within 10 minutes of submitting their application with coaching tips and insight into their personality.
“It’s the science of understanding people through language, and it’s very similar to what people are experiencing with the power of GPT,” Hyman said.
Customers include Medibank, Qantas, Woolworths and Suncorp.
If you look at your resume
Medibank Senior Talent Acquisition Executive Andrew Letchko said Sapia.ai was used by insurers as part of the hiring process for corporate roles and to address proximity bias when hiring talent internally. said to have used Proximity bias refers to the tendency of leaders to treat people who are physically closer to them more favorably. The switch to hybrid work caused by the pandemic has become a bigger issue for employers.
“This is one of the first things we ask candidates to fill out after they apply for a position at Medibank,” Rechko said.
“The talent market is highly competitive and this tool has been very helpful in identifying candidates who may have been overlooked based solely on resumes.
“It was also helpful when dealing with internal movements and proximity biases that may exist in the workplace.
“This tool helps our internal candidates show a side of themselves that they often don’t get to see in their day-to-day work, especially in a hybrid work environment.”
