A university graduate participates in an AI Interview. (Photo/Hongxing News)

For Luo Xinjing, a recent graduate from Sichuan International Studies University, facing an AI-generated interviewer on the screen is even more nerve-wracking than a real-life interview.

The AI interviewer asked, "You must put in significant effort to achieve better results. Can you share an experience where you did this, and what was the outcome?"

Unlike human interviewers, the AI interviewer's body language was somewhat mechanical. Luo often felt more nervous because her speech didn't flow smoothly while answering questions.

She explained that in a face-to-face interview, the interviewer can gauge confidence through facial expressions and interaction. However, the AI interviewer relies more on mechanical analysis and lacks human empathy.

Luo, who didn't have standout work experiences, felt she gave a weak answer to the AI interviewer's question.

"I talked about a leadership experience in a student club, but because I didn't organize my thoughts beforehand, the content was lacking, and the logic was somewhat chaotic. I believe this is why I didn't pass the interview."

Luo's AI interview experience isn't unique. Social media discussions about AI interviews, question banks, and complaints are trending. On Xiaohongshu or Little Red Book, there are over 50,000 posts with more than 50 million views on the topic.

With the development of AI technology in recent years, many internet companies, fast-moving consumer goods (FMCG) firms, and banks have gradually adopted AI interviews to improve recruitment efficiency. However, the interview experience has sparked considerable debate.

Some job seekers complained that the AI interviewer often repeatedly asks about topics already answered during follow-up questions. Another concern is whether information on personal privacy involving self-introduction and facial recognition will be deleted afterward.

Extensive use of AI interviews

Powered by artificial intelligence, the AI interviewer utilizes technologies such as speech recognition, natural language processing, and emotion and facial expression analysis to comprehensively analyze candidates' responses and behavior. This process aims to assess their abilities and potential.

"Over the past few years, AI interviews have advanced to generate authentic and objective portraits of job candidates in simulated interview settings, efficiently managing large-scale interview screenings. This capability has been particularly effective in campus recruitment and mass social recruitment," the head of the Business Research and Analysis division of the recruitment department at 58.com, one of China's major online marketplaces for classifieds, said to China Newsweek.

According to the media outlet, AI interviews leverage extensive big data to achieve precise candidate matching tailored to individual needs. This is particularly beneficial for mobile job markets like blue-collar and grey-collar workers, where approximately 300 million blue-collar workers change jobs three times annually, underscoring the potential of AI interviews.

Jin Ying, professor at the School of Politics and Public Administration at Southwest University of Political Science and Law, acknowledged the advantages of AI interviews. She said that the AI system's built-in competency models compensate for potential professional gaps in companies and that AI interviews reduce the energy, time, and costs involved in the interview process.

"AI is mainly used for structured interviews and does not include follow-up questions, but it can quickly capture relevant information with strong time control and high information acquisition rates,” Jin added. This type of interview also reduces human factors such as subjective impressions and bias, providing relatively objective assistance for personnel selection, according to Jin.

A White Paper on AI Applications in Corporate Human Resources released last year showed that AI-assisted interviews are already common in campus recruitment and mass interview scenarios. They are gradually evolving from "interview assistance" to "interview decision-making."

The white paper highlighted that Siemens receives over 150,000 resumes annually for campus recruitment. The AI interview system has reduced the interview load for business line leaders by 30 to 50 percent while increasing the recruitment success rate by 15 to 30 percent.

A product manager for AI interview-related products at recruitment portal Zhaopin told China Newsweek that AI interviews are currently used in industries such as internet, telecommunications, finance, real estate, biopharmaceuticals, manufacturing, energy and environmental protection, as well as in government and public institutions. Most users are large-scale enterprises.

A report on 2023 China Online Recruitment Development showed AI video interviews are used in 31.8 percent of scenarios. A market research firm, Resume Builde report, predicted that 43 percent of companies would incorporate AI in their recruitment processes by 2024.

Unresolved issues in AI interview

However, many job seekers have expressed feelings of anxiety and unease during AI interviews.

In Luo's opinion, job seekers are not machines, and using AI interviews to assess their suitability for a position can be too mechanical. Traditional face-to-face communication allows company leaders to exchange information and connect with candidates. Moreover, Luo noted that AI interviews can be cheated by reading from a script placed behind the camera.

Jin also noticed the drawbacks of AI interviews. For example, AI interviews are primarily used to assess knowledge and certain abilities but cannot cover all qualities or be applied to all candidates. While they meet some selection purposes, AI interviews cannot have the multiple values of traditional ones, especially emotional communication, making it hard for employers to show human care and for candidates to express emotions. AI still needs improvement in detecting fraudulent or performative interviews.

Given this, Jin believes AI interviews in the recruitment process could introduce the company and job responsibilities to candidates, conduct preliminary screenings of knowledge, language skills, and intelligence to reduce the workload of early interviews and collect candidate data for subsequent analysis and management.

Additionally, AI interviews collect personal information such as identity details, educational background, facial expressions, voice, and video, posing privacy risks.

According to Jin, companies need clear guidelines on information collection, storage, and usage rights, as well as candidates' rights to choose and be informed. Legal clarity is needed to link these issues with privacy protection laws and introduce relevant regulations. Meanwhile, the legality of data collection and usage should be tied to the company's credit and qualifications to increase responsibility and the cost of errors.