Using AI, ML, and advanced NLP, Computer Vision techniques.
Artificial Intelligence (AI) is intelligence demonstrated by machines. The primary objective behind its finding was to enable computers to perform human mental functions, including learning, solving logical problems, and communicating using natural languages with humans. AI is capable of accessing and processing boundless amounts of data, a million times faster than the smartest human to potentially solve every problem, including how to run an economy more successfully and win every possible battle. Hence, the impact of AI on society will be enormous. But can we use AI for selecting the right candidate for a job role out of huge lots of applications received? Can a robot take an interview in place of a human being? Everyone would love the idea of a system doing their mundane tasks, and on the verge of the fourth industrial revolution, there stands a good chance of AI technology actually augmenting such human capabilities. For instance, automating screening, onboarding, hiring, and recruiting processes in companies. Automated systems or models can help humans to get results in a cost-effective and timely manner. Still, the danger of artificial intelligence can never be ignored; hence any plan framed should only be done after a series of meticulous research or study.
The recruitment process is quite tiring, especially when a single person needs to sort out things and select the correct resume or candidate from thousands of applications. Many companies face a similar problem and sometimes also struggle to complete the whole process of interviewing candidates. Intentionally or unintentionally, they keep on delaying the process of one of the recruitment rounds, mainly the final HR round. The main reason was the unavailability of a human resource for conducting the round. This is where automation can help the companies, like using ATS systems during the screening process and advanced NLPs for framing the questions rounds like the HR Round. Though the former idea looks impressive, it is not only one solution to all the problems faced by companies during the recruitment process. If the resume does not satisfy the same ATS system’s requirements, it penalizes the input or resume, and hence it gets rejected. The ATS system is not a human; after all, it can only parse the resumes during the screening process. It cannot sit with applicants to judge their emotional quotient, personality traits, etc.
Natural Language Processing has already helped many companies to achieve various solutions in the form of Sentiment Analysis, Email Filtering, Voice Assistants/Recognition, Machine Translation, and a few others. However, the systems imitating humans, like forming general sets of questions using advanced NLP techniques and solving lexical and syntactic ambiguities, are not fully built yet. Many companies have tried to make use of the advanced NLP techniques, but the results did not come out to be 100% effective. One of the reasons is that the rules in natural language processing applications to the formal language mostly follow a specific set of words contained in the vocabulary and then sort out the problems. But in real life, the causal texts imputed by humans contain a series of grammatical errors, punctuation errors, no use of capitals for initials, missing abbreviations, use of letters in place of numbers, or cases like different dialects, accents, contexts, etc. It is fair to say that acquiring knowledge of the world is difficult for systems or models. Furthermore, it is hard for systems to sense personality traits like honesty, punctuality, interpersonal sensitivity, and other behavioral aspects. Coupled with these, the unavailability of real-time datasets and implement the solution using existing AI/ML models makes the task more challenging. It is tough for humans to gather huge chunks of data. Hence, for general practices, previous representations of words gathered after computation from past datasets are used, and then hyper-parameter tuning is done.
A couple of years back, Amazon Inc faced a backslash when many of its machine-learning specialists identified a major flaw with their AI recruiting tool. It was found that the tool was gender-biased as the recruiting engine did not recruit women applicants. It simply rejected the applications with resumes that had words like “women,” “lady,” or “female.” As a result, the company banned the project as they lost their hopes.
Hence, in order to ease the interview process, it is necessary to investigate various aspects and think of more efficient strategies for matching job descriptions and filtering out ideal candidates.
Manual filtering of the best fit resumes out of the massive number of applications received by a company through the internet takes a lot of time. It becomes practically impossible for companies to select ideal ones in such a short span. Furthermore, this method of screening and filtering out resumes by using algorithms does not seem to be a fair approach. A number of suitable candidates do not even get a chance at times. Hence, the future prospects should be to mine the candidate’s profile data on various social platforms like GitHub, LinkedIn, Facebook, etc. Gathering the social behavior information coupled with their resume content can help the companies to filter out ideal candidates in a more efficient way (Chirag, Harsh, Gurneet, & Indrajeet, 2020).
The suggested approaches in studies are acquired by integrating two types of information: context knowledge and ontology of ideas, knowledge of an interesting topic, and incorporating them into shallow natural language processing (Omar, Shahida, & Norita, 2011).
As firms are increasingly implementing algorithmic decision tools to reduce human bias, save money, and automate procedures, our analysis reveals that algorithms are not neutral or devoid of biases just because a computer generated a certain choice. Humans should continue to play a significant and crucial role in the proper governance of algorithmic decision-making (Alina & Marius, 2020).
AI application in the recruiting process is still in its early stages in India, and it is restricted to specific aspects of the recruitment process. Companies that use AI apps in the recruiting process are either technologically creative or very large enterprises with a lot of room for expansion. The scope of AI application in the recruiting process is quite broad, and it still has a long way to go. According to studies, the use of AI in the recruiting process benefits not only the recruiters but also the candidates. On the other hand, fast advancements in information technology and processing capacity, together with falling technological costs, will lead to the growing use of AI and ML (Rahul & Usha, 2021).
Based on current data sets and behavioral patterns, artificial intelligence enables robots to make better judgments than people. As a result of this shift, machines have taken over all physical labor, forcing HR professionals to take on more strategic duties. It is critical for businesses and professionals to grasp how this technology works and its role in various HRM activities (Isha & Mohit, 2020).
Transformation of Human Resource Management, specifically the recruitment processes by making use of AI/ML-based applications
The nature of the study here is loosely structured, and the observations of various aspects determine the accuracy of the findings of the research. For this study, a number of research questions can be framed
- Will AI completely replace all types of human jobs?
- Is there any possibility of building advanced NLP models that can cater to the recruitment needs of companies?
- Will systems adopt biases like affinity bias, halo effect, beauty bias, gender bias, horn effect, etc.?
- Can the model gain enough potential to make judgments based on human emotions, traits, or behaviors?
This study frames Exploratory research as very few studies or researches have been conducted considering the advanced aspect. Sometimes this research is informal and unstructured. It serves as a tool for initial research that provides a hypothetical or theoretical idea of the research problem. It will not offer concrete solutions for the research problem. This research is conducted in order to determine the nature of the problem and helps the researcher to develop a better understanding of the problem. Exploratory research is flexible and provides the initial groundwork for future research. The study looked for evidence to support the problems with recent practitioner sources, and secondary information related to the companies.
Timeline & Budget Table
- Artificial Intelligence or Machine Learning is not the only savior. Thus, the idea of entirely relying on them is not correct. Research has shown that it would further take a lot of years for AI to completely rule and replace all human tasks. As of now, it can only be used as something to augment human capabilities
- Many cases happened in recent years where the failure of automated recruitment resulted in a setback for the companies, so suitable models or algorithms should be implemented for the process.