Your interviewer may be an AI

When we, carbon-based life forms, encounter silicon-based AI interviewers, the recruitment field is experiencing unprecedented changes. The application of AI technology has brought both efficiency and innovation to job seekers and companies, but it has also triggered in-depth thinking about security, fairness, and privacy protection. In this new era full of opportunities and challenges, how can we grasp the pulse of AI recruitment and achieve harmonious coexistence between humans and machines?

When carbon-based meets silicon-based, we can explore this topic from multiple angles.

From the perspective of life forms, research on silicon-based life forms has shown that the similarity in chemical properties between silicon and carbon makes it possible for silicon-based life to improve efficiency in AI interview scenarios.

The essence of an enterprise is efficiency and innovation. This new change in AI interviewers not only broadens our understanding of the diversity of artificial intelligence, but also provides a data foundation for the business world to explore more AI big model scenarios.

This year, the traditional "Golden March and Silver April" recruitment season has begun. Workers who are preparing to change jobs are eager to try, and tens of millions of college graduates are on their way.

Value Planet found that one significant change this year is the large-scale application of AI technology, such as graduates using AI to write resumes, AI simulated interviews, and AI recommending jobs; HRs also improve work efficiency through AI screening, AI research, AI interviews, etc. Both parties are trying to find the "most suitable person" in the vast sea of ​​​​people.

There are pros and cons to everything.

The security of AI recruitment has also raised a series of issues. In January this year, according to a report by foreign technology media 404Media, there was a vulnerability in the AI ​​recruitment system used by the global fast food chain KFC. Hackers could not only steal job applicants' information, but also manipulate the AI ​​system to reject fast food restaurant applicants, and even hire or fire them.

The Beijing News pointed out that AI recruitment may have the problem of "employment discrimination". If a corporate recruiter who already has biases such as gender, region, and age uses this algorithm system to screen resumes, then objectively a "discriminatory information cocoon" is created in the process. In the future, the resumes that can be retrieved and pushed by the algorithm are all those that have been filtered through bias. This makes the "discrimination" problem, which was originally relatively invisible, even more difficult to detect.

In response to new problems that may arise in AI recruitment, the country has successively issued the "Internet Information Service Algorithm Recommendation Management Regulations" and the "Guiding Opinions on Strengthening Comprehensive Governance of Internet Information Service Algorithms", and continued to carry out special operations to thoroughly investigate and rectify algorithm abuse in various companies.

While supervision is being strengthened, technical means are also being strengthened. For example, in the review of the 2021 AI technology support special project, Shanghai has specially deployed a special topic on the construction of an AI social experiment system, emphasizing the need to scientifically evaluate the comprehensive social impact of AI, effectively promote the construction of AI application standards, and assist in supporting public governance decisions on AI applications.

Thanks to the efforts of all parties, AI recruitment has penetrated into every aspect of recruitment and job hunting. Mainstream recruitment platforms such as BOSS Direct, Liepin, and Lagou, and technology providers such as Beisen and Haina AI have launched AI recruitment products, covering job recommendations, AI recruitment assistants, AI interviewers, AI coaches and other services.

1. Whether you can join the company or not is decided by the AI ​​interviewer

Fanfan, a post-00 master's graduate, is actively preparing for this year's spring recruitment. In the past three months, he has had more than 20 interviews of varying sizes.

Unlike their seniors, these post-00s "new generation" are more adept at using various technical software. "My resume was optimized using AI, and my job search direction was also determined by AI testing. Because I majored in logistics, but didn't want to work in related fields, but was interested in new media operations, I took several personality tests, career tests, and other tests, which finally strengthened my determination to find a related job," Fanfan said.

Before the on-site interview, Fanfan participated in several AI interviews. When he turned on the phone screen, he saw an AI interviewer in front of him. The AI ​​interviewer would ask questions like "What did you learn from the self-media operation internship you mentioned in your resume? How was the operation effect?" and so on. Fanfan also summed up his own answer methodology, "AI is not that smart after all. As long as the answer contains some keyword information, it can get a higher score. I usually reply to AI with three key points: what to do, how to do it, and methodology. I also have to keep looking at the screen without dodging my eyes. This way, I can get a good score for the first interview."

The so-called AI interview is to integrate artificial intelligence technology with real interview scenarios, analyze the candidates' words, speaking speed, movements, expressions, etc., and analyze the candidates through data quantification.

A month ago, Hangzhou Normal University and Zhejiang University of Technology held offline spring recruitment conferences at the same time. The on-site "AI resume clinic" and "AI interview" were welcomed by graduates. Graduate Xiao Xu said: "I simulated the interview scene with the AI ​​interviewer. After a few minutes of answering questions, the system intelligently evaluated the interview performance in terms of behavioral ability, growth characteristics, cognitive ability, etc., which also made it convenient for me to find my own problems and conduct targeted training later." The fastest AI interview, selecting a position, scanning the face to submit, and confirming the information, can complete a simple job search within 15 seconds.

According to the "2024 Niuke Intelligent Manufacturing Campus Recruitment White Paper", 53.5% of students interested in intelligent manufacturing have participated in AI interviews, and 77.7% of students who have participated in AI interviews expressed satisfaction.

Image source: iResearch

From the perspective of enterprises, AI interviews can help employers save labor costs and improve analysis efficiency. According to data released by Liepin, its AI product Doris can complete the assessment of 40,000 people in 8 hours, which can reduce the recruitment costs of enterprises by 80%.

AI interviews have become a "must-have" for many large companies. As early as 2019, Unilever, L'Oréal, and Coca-Cola have adopted AI for interviews. Bank of China, China Merchants Bank, China Mobile, etc. in China have also started AI interviews in recent years.

On the Zhihu platform, the topic of "AI Interview" has generated 1.36 million views, and experience sharing posts such as "AI Interview Scoring Points", "AI Interview Summary and Notes", and "AI Interview Answering Skills and Methods" have all received high views.

Source: Provided by the interviewee

Fu Dan, who just participated in an interview at China Mobile's Sichuan branch, said that it was her first AI interview and she had no time to introduce herself before she entered the question-answering phase. In the 13 questions, most of them were about interpersonal relationships in the workplace. "I was totally unprepared and my answers were fragmented. Fortunately, I was notified that I passed the interview later," said Fu Dan.

2. AI recruitment enters a stable development period

This year's government work report proposed to deepen the research and development and application of big data, artificial intelligence, etc., and launch the "artificial intelligence +" action.

The AI ​​interview is one of the practical application scenarios of "artificial intelligence + recruitment". It adopts a combination of AI and manual interviews. AI interviewers first conduct large-scale screening of job applicants, and then real interviewers make the final review, thereby greatly improving efficiency.

Dai Kebin, Chairman and CEO of Tongdao Liepin Group, said, "In the face of changes in organizational forms and talent standards, AI will play an important role in organizations, and every company should have an HR digital assistant in the future." In his view, the core of AI smart interviews lies in truly understanding the essence of interviews, and then empowering AI with scientific and standardized offline interview processes and skills, so that it has the talent identification capabilities of senior interviewers, rather than just providing a simple technical tool.

As Dai Kebin said, the application of AI technology in the recruitment process has entered a period of stable development, and many leading platforms have launched related AI products. For example, BOSS Zhipin mentioned in its 2023 annual report that the company's independently developed vertical language model "Nanbei Pavilion" for the recruitment industry has been filed through the "Interim Measures for the Management of Generative Artificial Intelligence Services". Liepin mentioned in its third-quarter 2023 financial report that the company has launched "Super Chat", an intelligent contact tool to help companies recall suitable talents. In March of this year, Duomian·Doris was launched. In addition, Lagou has also launched the large model HRMind, Zhaopin launched AI Yimian, 58.com launched AI Recruitment Superman, and so on.

An industry insider revealed that the main application scenarios of AI interviews at this stage are reflected in campus recruitment, blue-collar workers, sales and customer service, and foreign language positions, while for positions that require more creative thinking, such as programmers, designers, editors and reporters, AI interviews perform mediocrely. For example, Sungrow, a leading company in the solar energy storage industry, decided last year to use Niuke AI interviews for all English interviews for management trainee positions.

But just as "a coin has two sides", there are also two sides to a thing. While enjoying the efficiency and convenience of AI, the fairness of AI interviews has also caused concerns. In extreme cases, a corporate recruiter who already has biases such as gender, region, and age is likely to train an AI with "employment discrimination", and using this algorithm for recruitment will inevitably affect the fairness of recruitment.
In such a situation, AI has become a tool to amplify prejudice and has caused wider "workplace harm."

Therefore, relevant people suggest that the "red line" of fairness and justice should be included in the AI ​​algorithm rules. For example, in the interaction with users, the use of suspected discriminatory behaviors should be defined, monitored and restricted. Or in the recommendation rules, emphasis should be placed on achieving a balance in the proportion of gender, education level, region, etc., to ensure that every job seeker has the opportunity to be seen to the greatest extent possible.

On the other hand, with the trend of "examination-oriented" recruitment, interviewees who are good at making strategies are more likely to get high scores, while some applicants may be eliminated due to "insufficient preparation" or even "social phobia", which may hinder the original intention of the recruiting unit in identifying and selecting talents.

At present, it seems that AI interviews still need to accumulate a large amount of data to be improved.

3. Will AI replace HR?

Value Planet believes that there is a high possibility that AI interviewers will become a future trend.

AI interview systems can adapt to the employment needs of enterprises more quickly, but they cannot completely replace HR work. This is because the core work of HR is dealing with people. The personality and career needs of interviewees vary from person to person. Only in-depth communication between people can implicit information be transmitted.

Although AI interviewers play an increasingly important role in the recruitment process, corporate employment needs still require heavy participation and rational judgment from professional HR personnel.

At the end of last year, at the "Artificial Intelligence Safety Summit" held in the UK, Elon Musk, founder and CEO of Tesla and co-founder of OpenAI, said that "artificial intelligence will enable humans to reach the point where 'there is no need to work'."

Judging from the wide application of AI in the field of recruitment, it has indeed greatly reduced the workload of humans.

Last year, 58.com launched an intelligent AI interview tool - AI Recruitment Superman. At the product launch conference, the relevant person in charge introduced its usage scenarios. If you enter "I want to recruit an accountant" in AI Recruitment Superman, you can get relevant job interview questions and provide relevant skills and feedback references for HR reference and use; if you enter "Help me write a waiter recruitment notice", the system can combine the characteristics of the industry and the issues that job seekers are concerned about to generate a professional recruitment notice.

It seems that AI recruitment can even replace real HR to some extent, and some practitioners have even felt the crisis of their careers.

Ma Li, a user of Duomian Doris, said, "Doris' current level is roughly equivalent to that of an HR with two or three years of experience, but according to the current progress of technological development, it is estimated that she will soon have five or ten years of experience. By then, there may really be no need for real people."

The company where Ma Li works needs to recruit hundreds of telephone customer service staff every year due to the high staff turnover. "It takes about 15 minutes to interview a candidate according to the process and established questions, and then decide whether to proceed to the next step of communication based on the internal scoring. Now, an AI interviewer can complete an interview in a few minutes, and is available 24 hours a day, 7 days a week. In comparison, the efficiency of real people is really too low," said Ma Li.

A survey conducted by 36Kr's "Windvance" in May last year showed that 75% of the HRs interviewed believed that their jobs would be partially replaced by AI, and 3% of the HRs interviewed believed that AI would completely replace their jobs.

But some people are not worried about the impact of AI at all. Rebecca, an HR of a large Internet company, revealed that although she also uses AI tools as an aid, she is not worried about losing her job to AI. This is because AI cannot ask many logical questions during the interview process, and many people need to have a real chat before they can confirm whether they are suitable. AI cannot judge the "aura" at all.

She gave an example, "When I interview candidates for high-end positions, I will ask them questions such as their career plans and career ideals to judge their enthusiasm for the position. Some people really don't change jobs for the sake of income. Their real needs may be varied, such as disagreeing with the values ​​and corporate culture of the original company, or the original job does not bring a sense of accomplishment, or the original job does not bring professional skills growth, etc. In the face of these uncertain needs, our HR needs to make screening and judgments. Only by understanding their real needs can we put forward attractive conditions in the next communication and discover and retain suitable high-end talents."

Author: Tang Fei, Editor: Weber

Source: ValuePlanet (ID: ValuePlanet)