Saurabh said recent advancements in machine learning mean that new AVI technologies can not only record video but effectively assess and analyze interviewees via computer vision, voice analytics and natural language processing to measure soft skills such as professionalism, sociability, positive attitude and communication skills, among other metrics. He added that AI can also help rank applicants to make it easier for admissions officials to shortlist large pools of candidates for the first round of human interviews.
According to Saurabh, roughly 70 percent of the top universities in Singapore, including the National University of Singapore (NUS) and Nanyang Technological University (NTU), now use Interviewer.AI for admissions interviews as the technology gains more interest in higher education and the company looks to penetrate other markets moving forward.
“With the pandemic still raging in different parts of the world, in-person campus interviews for thousands of applicants are a challenge and can trigger a super spreader event. Universities can leverage AVIs to screen applicants in addition to other relevant parts of the application process, to not only save time and costs for both universities and the students, but also provide a great experience to every applicant,” he wrote in an email to Government Technology, noting that the admissions process is an early use case for AI-driven AVI technology.
Aside from streamlining the admissions process, Saurabh said, higher education institutions may also increasingly look to AVI technology for career services on campus, which help students find jobs relating to their degree programs following graduation.
He said another use case for the emerging technology is to help universities coach final-year students for their first job interviews, and employers to score resumes for job experience and academic qualifications, among other uses.
“Career services [at universities], which usually consists of less than 10 faculty members, is required to coach hundreds of students from the same batch and target 100 percent placement,” Saurabh noted in an email. “This can be an arduous task, especially when most of the companies that youngsters in college want to join are multinational companies that typically have an asynchronous video interview before a human recruiter can evaluate the candidates.”
According to a 2020 essay by Alan Jones, Suzan Harkness and Nathan Mondragon for the nonprofit Educause that examines the efficacy of AVI tech tools for processes such as these, advancements in AI for asynchronous interviews could create a “shift to a new paradigm for interviewing and hiring,” as well as for career guidance in higher education. The essay cites 2018 survey results from the National Association of Colleges and Employers that found many employers are seeking new methods for screening and filtering applicants beyond degree obtained, schools attended and GPAs, which can prove less predictive of success at work than general mental ability and soft skills.
“Employers wanting to cast a wider net to increase the diversity and inclusion of hiring pools can use video interviewing technology to more easily reach larger applicant pools of diverse candidates,” Educause wrote. “Breaking free of campus-based career fairs allows employers to recruit nationally and internationally without the need to physically be on multiple campuses or make choices based on recruiting and travel budgets, college size or ranking.”
Efficiencies aside, Educause urged caution and the need to "not be too quick to cede power to AI or give too much attribution to the power, scale and ability of machine learning and AI in its present form.” It argued that historical data ingested through machine learning is often biased with respect to gender, race, ethnicity or social class.
Despite these concerns relating to bias in AI, Saurabh believes recent improvements in AVI and AI technology could still prove particularly useful to screening large pools of applicants without inherent bias.
“The data aggregated using AVIs can bring in a lot of objectivity and data-driven insights, which are typically lost in an environment where thousands of applicants are screened by human personnel whose individual experience, and skills in finding the right candidate, may vary, resulting in conscious and unconscious bias,” he wrote in an email. “Using AVIs, companies like Interviewer.AI can not only present data and insights to the various stakeholders, but they can also actually get better with time by incorporating machine learning techniques based on the success of screened candidates’ performance over time.”
Touching on solutions, the Educause essay notes that vendors involved in AVI tech development have worked in recent years on procedures to identify biased variables that were built into data sets and algorithms, and removing features that used those variables. Some hire third-party auditors to help mitigate the risk as well.
Still, according to Educause, data sets that go into algorithms like those used for AI-driven AVIs can mirror human judgments that are inherently subjective.
“Although attempts to democratize experiences and clean the data might instill more trust from the public, there are some broader aspects to machine learning that need to be addressed as we scale AI more broadly and establish reliability and validity,” Educause wrote.
