More and more companies are using artificial intelligence to recruit and hire new employees, artificial intelligence can be considered Almost any stage in the recruitment process. Covid-19 has driven new demand for these technologies.Both Curious stuff with HireCompanies that specialize in artificial intelligence interviews report that their business has surged during the pandemic.
However, most job searches start with a simple search.Job seekers turn to platforms such as LinkedIn, monster, Or Post Code Recruitment, Where they can upload resumes, browse job information and apply for vacancies.
The goal of these sites is to match qualified candidates with available positions. In order to organize all these vacancies and candidates, many platforms have adopted artificial intelligence-driven recommendation algorithms. Algorithms, sometimes called matching engines, process information from job applicants and employers to compile a recommendation list for each person.
“You usually hear anecdotes about recruiters taking six seconds to look at your resume, right?” said Derek Kan, vice president of product management at Monster. “When we look at the recommendation engine we built, you can reduce the time to a few milliseconds.”
Most matching engines are optimized to generate applications, saying John Jessing, Former vice president of product management at LinkedIn. Recommendations of these systems are based on three types of data: information directly provided by users to the platform; data assigned to users based on other people with similar skills, experience, and interests; and behavioral data, such as the frequency of users responding to messages or interacting with job postings.
In the LinkedIn case, these algorithms exclude a person’s name, age, gender, and ethnicity, because including these characteristics may lead to bias in the automated process. But Jersin’s team found that even so, the service’s algorithm can still detect the behavior patterns of groups with specific gender identities.
For example, although men are more likely to apply for jobs that require work experience beyond their qualifications, women often only choose jobs whose qualifications match the requirements of the position. The algorithm explains this change in behavior and adjusts its recommendations in ways that are unintentionally detrimental to women.
“For example, you might recommend more senior positions to one group of people than another group, even if they have the same level of qualifications,” Jersin said. “These people may not be exposed to the same opportunities. This is indeed the impact we are talking about here.”
Compared with women, men include more skills in their resumes, but are less proficient than women, and they generally interact more actively with recruiters on the platform.
To solve these problems, Jersin and his LinkedIn team Built a new artificial intelligence Aims to produce more representative results and deploy it in 2018. It is essentially a separate algorithm designed to counteract recommendations that are biased towards specific groups. The new AI ensures that the recommendation system includes even distribution of transgender users before referencing matches curated by the original engine.
Naoto Kan said that Monster will list 5 to 6 million jobs at any given time, and it also incorporates behavioral data into its recommendations, but it does not correct prejudices like LinkedIn does. Instead, the marketing team focused on getting users from different backgrounds to sign up for the service, and then the company relied on employer reports and told Monster whether it passed a representative set of candidates.
Irina NovoseelskiThe CEO of CareerBuilder stated that she is focused on using the data collected by the service to teach employers how to eliminate bias in recruitment information. For example, “When candidates read job descriptions with the word’rock star’, the percentage of women applying is significantly lower,” she said.