Recently pull the hook. Com announced the launch of a brand-new enterprise and recruiter identity audit mechanism, and introduced Baidu AI collaboration section, and introduced intelligent methods such as face recognition to audit and verify the HR identity and qualification of enterprises. In the future, artificial intelligence technology will be used to verify business licenses, work permits, employment permits, and risk control of resumes and chat records.
Since AlphaGo defeated the world champion of Go last year, the commercial application speed of AI technology has obviously accelerated, and it has begun to replace labor in transportation, family service, medical care, commerce, recruitment and other fields. Some practitioners feel the pressure while enjoying the convenience brought by AI technology.
The recovery speed of robot screen "seconds kill" human beings, but it is not flexible enough.
A few months ago, a "man-machine war" sponsored by the recruitment industry gave the answer with scores. The challengers are five senior HR and headhunters from Internet companies, and the two sides should quickly select the 10 resume that best meets the requirements of the recruitment position from 37 million resumes.
This is the matching between job requirements and candidates, including technical posts and product posts. After completing the selection and matching of resumes, the jury needs to score the contestants from six dimensions: function, skill, industry, salary, education and regional matching. The person with the highest total score (out of 25) is considered to have won the game.
As the blue screen lights up, the ratio of human to AI robot is 18.96: 18.60. The results show that the AI robot only needs 0.0 152 seconds to complete the whole journey, which is 63,882 times the average speed of human beings. The matching efficiency of robots is higher than that of humans in job matching and region matching. In terms of skill collocation, the two are tied.
Although it lost to humans with a weak score of 0.36, AI robot still exceeded the expectations of Dai Kebin, founder and CEO of Hunting Net, in terms of matching people and understanding people. "In the task of resume search, areas, salaries and other aspects are relatively simple and direct conditions, so the algorithm can be realized through simple logic without making mistakes; From the perspective of industry background and skill requirements, the algorithm has been able to make more accurate understanding and similarity judgment by using technologies such as neural network and natural language processing. " Chief data officer Yi Shan, the designer and hunter of this Bole robot, told China Youth Daily and Zhongqing Online reporter that at present, AI robots have been able to better understand most explicit needs, such as functions, skills, salary, education, and region. The matching level of algorithm can be comparable to that of professional recruiters; However, under the implicit conditions that culture, values and temperament need face-to-face communication, algorithms cannot replace human communication and judgment.
In the results, there is the biggest gap between robots and humans in the degree of academic matching, and the main reason for this gap is that robots can't identify what kind of academic qualifications "upgrading undergraduate" belongs to. This also shows that the flexibility of robot thinking is limited. In this regard, Yi Shan explained that the job requirements of "bachelor degree or above" are judged to meet the requirements when designing the robot algorithm. But in fact, in the eyes of many recruiters (especially high-end headhunters), undergraduate courses are not as good as undergraduate courses. Therefore, this screening result caused different opinions of several judges present. "Robots are not enough in personalized selection of talent soft indicators according to corporate and HR preferences."
Zhou, a member of the jury and an expert in recruitment of Alibaba Entertainment, believes that the gap between robots and humans is almost negligible when selecting people from a large number of resumes on a large scale, and AI robots can improve the overall recruitment efficiency.
It is nothing new that robots screen resumes much faster than humans. According to reports, in March this year, in an industry competition held by SourceCon, a well-known headhunting company in North America, a robot "Brilent" based on artificial intelligence to screen and rate job seekers took only 3.2 seconds to screen out the suitable candidates from 5,500 resumes, ranking third among the contestants in accuracy. Based on the data structure processing and detailed field matching experience accumulated by members on Facebook, this team uses AI technology to sort job seekers by "person-post matching", freeing HR from mechanical and cumbersome resume screening, and can focus more on the follow-up interview selection process.
High-precision job matching: let artificial intelligence learn how HR does recruitment.
In June this year, Dai Kebin announced that hunting should further improve recruitment efficiency and enrich recruitment ecology through the exploration of big data and artificial intelligence; 12 In September, Kai-Fu Lee, Chairman and CEO of innovation works, said at the "20 17 China Artificial Intelligence Summit" that sufficient data and accurate scenes are the prerequisites for artificial intelligence to truly replace manpower.
From simple job information listing and classification to job matching system based on big data mining, in recent years, many online recruitment enterprises in the whole industry have formed their own "talent pool" through data accumulation. On this basis, matching people and posts according to resume information or job requirements has become the main application of artificial intelligence technology in the recruitment field at this stage.
As a start-up Internet company specializing in providing mobile recruitment services for enterprises, Wang Daozhi, CEO of Qian Xun Mobile Recruitment, once pointed out in an article that whether the text matching technology based on JD (job description) and resume can reach the level of effective selection and whether the output of the system (machine learning process) can approximate the result of manual selection can measure the effectiveness of AI. This process includes two dimensions: the matching of resume information and job requirements, and the matching of candidates and corporate jobs.
The application of "AI+ Recruitment" in hunting began on 20 14. According to Yi Shan, in the process of R&D and exploration, they found that it is a challenge to apply artificial intelligence to the recruitment industry to judge whether it can make high-precision recommendation based on user behavior data, position and resume content, whether it has a detailed understanding of job requirements and industry fields, and whether it can achieve personalized job recommendation for job-hoppers across occupations and industries. It is understood that at present, most recruitment websites can use the data capture and semantic analysis of artificial intelligence to achieve the matching of the first dimension, while the accurate matching of the second dimension puts forward higher requirements for personalized recommendation of artificial intelligence.
"Machine screening first needs to accumulate knowledge and experience and be able to interpret semantics; Secondly, we must learn the behavior of HR and headhunters and know how they match people and posts. " In Yi Shan's view, the latter is more critical, namely, the use of depth portraits, semantic matching and personalized recommendation based on HR preference.
In order to make artificial intelligence "think" like HR, the single art leading team with 17 years of experience in data mining and system research and development of domestic and foreign enterprises requires machines to actively learn from humans. They designed a set of algorithm system, which was open to the HR of the client company and allowed HR to give feedback on the "resume recommended by the machine". When the feedback behavior and data accumulate to a certain extent, the machine can understand the preferences of different HR from the selection differences and form a matching model that can handle thousands of different industries and functions. Even for the same position, different recommendation results may be given.
"Now my robot Bole is already doing HR." According to Yi Shan, at present, the business volume recommended by the algorithm far exceeds 50% of the total business volume, and the accuracy is the same as that of general headhunters.
This intelligent recommendation algorithm is called "intelligent system" by them. "Let the robot learn how to recruit HR instead of telling him' how to do this and how to do that'." Single art emphasizes.
Open the information gap between enterprises and job seekers
Artificial intelligence is becoming more and more popular in the recruitment industry, behind which is the growing demand for recruitment business.
According to the statistics of iResearch, it is estimated that by 20 18, the number of small and medium-sized enterprises in China will exceed 87 million, and the number of job seekers is expected to exceed 654.38+600 million. Searching a large number of resumes and screening potential candidates to match positions in various industries (especially non-high-end positions) has become the most repetitive job in headhunting and HR recruitment.
According to media reports, from July 20 16 to June 20 17, the application of AI has gradually spread to the recruitment process in 68 countries around the world. In the past year or so, Unilever has tried to hire artificial intelligence to recruit employees in North America, including using algorithms to screen resumes, game tests and face recognition, even without the participation of human interviewers. In China, as of July this year, there are no fewer than 10 start-ups claiming to be artificial intelligence+recruitment leaders, and efforts are made to solve the problems of high recruitment labor cost, low actual conversion rate and poor information between recruitment and job hunting through technology.
Niuzhi, an internet intelligent recruitment platform established on 20 16, adopts the following methods: accurately and comprehensively match the positions and resumes of enterprises (especially small and medium-sized enterprises) through resume decomposition and personalized recommendation; Teamable, a start-up recruitment platform, uses AI algorithm to mine candidates' social network data, trying to cut through social records and create an accurate closed loop of talent recommendation; The mini-school, which is perpendicular to the field of campus recruitment, also designs an intelligent matching model through data mining and AI algorithm, automatically screens and recommends resumes for different enterprises, and provides suggestions for job seekers with 0~3 years of professional experience.
However, Yi Shan has always stressed that in the recruitment industry, artificial intelligence is only a tool, which cannot replace human beings, but "helps human beings to make more accurate and well-founded judgments, so that headhunters and HR can engage in more valuable and creative work." To some extent, the application of artificial intelligence in the recruitment field is more likely to be promoted in low-end talents or job recruitment with low requirements for industry experience.
A headhunting consultant who specializes in high-end talents in the consumer goods industry admitted to the China Youth Daily and Zhongqing Online reporter that although it is necessary to frequently change the resume of the keyword search talent pool and make more than a dozen phone calls every day to contact candidates, the professional communication experience with distinctive social attributes such as improvisation required in this process is difficult for artificial intelligence to achieve.
"The more high-end talents and important executive positions, the more cautious the HR of enterprises, and the more professional headhunters are needed." Wang Guangyuan, general manager of Baiyi Century Management Consulting Co., Ltd. also suggested that candidates for functional positions such as product managers should also be considered through soft indicators such as product design concepts. "This is the weakness of the machine." Dai Kebin pointed out that "artificial intelligence can't replace headhunters immediately, and the lack of data between supply and demand is the root cause." He mentioned that job matching puts high demands on the job seeker data held by the recruitment platform and the demand description provided by the enterprise. Coupled with dynamic and flexible recruitment, while improving smart products, "human factors play an irreplaceable role in artificial intelligence recruitment."
In the current autumn recruitment season, in the face of a large number of fresh graduates pouring into the job market and the imbalance between supply and demand in the campus recruitment market, Yi Shan expects that in the future, AI technology can tap the personalized needs of enterprises by analyzing the supply and demand data of market positions, give appropriate employment guidance to fresh graduates in advance, broaden their horizons and choices, and make the needs of enterprises and fresh graduates match more efficiently.