Description:FORM 2
THE PATENTS ACT, 1970 (39 of 1970)
&
THE PATENT RULES, 2003
Complete Specification
(See section 10 and rule 13)
1. Title of the Invention: ARTIFICIAL INTELLIGENCE TECHNIQUES IN HUMAN RESOURCE MANAGEMENT
2. Applicants
Name
Nationality
Address
Dr. V. Gowrishankkar
Indian
Assistant Professor, Department of School of Management, Sri Krishna College of Technology, kovaipudur, Coimbatore-641042, Tamil Nādu, India.
Dr. Augustine George
Indian
Professor, Department of Computer Science, Kristu Jayanti College (Autonomous), K. Narayanapura, Kothanur, Bengaluru-560077, Karnataka, India.
Mr. Lijo P. Thomas
Indian
Assistant Professor, Department of Computer Science, Kristu Jayanti College (Autonomous), K. Narayanapura, Kothanur, Bengaluru-560077, Karnataka, India.
Mr. Jais V. Thomas
Indian
Assistant Professor, Department of Commerce and Management, Kristu Jayanti College (Autonomous), K. Narayanapura, Kothanur, Bengaluru-560077, Karnataka, India.
Mr. Emmanuel P. J
Indian
Assistant Professor, Department of Psychology, Kristu Jayanti College (Autonomous), K. Narayanapura, Kothanur, Bengaluru-560077, Karnataka, India.
Mr. Joshy Mathew
Indian
Assistant Professor, Department of English, Kristu Jayanti College (Autonomous), K. Narayanapura, Kothanur, Bengaluru-560077, Karnataka, India.
Mr. Deepu Joy
Indian
Assistant Professor, Department of Life Science, Kristu Jayanti College (Autonomous), K. Narayanapura, Kothanur, Bengaluru-560077, Karnataka, India.
Mr. A. Sevuga Pandian
Indian
Associate Professor & Head, Department of Computer Science, Kristu Jayanti College (Autonomous), K. Narayanapura, Kothanur, Bengaluru-560077, Karnataka, India.
3. Preamble to the Description:
The following specification particularly describes the invention and the manner in which it is to be performed.
4. DESCRIPTION
FIELD OF THE INVENTION
The present invention relates to the field of Human Resource Management. The main aim of this invention is to show the Artificial Intelligence techniques in Human Resource Management.
BACKGROUND OF THE INVENTION
The speed with which the business rhetoric in management moved from big data (BD) to machine learning (ML) to artificial intelligence (AI) is staggering. The match between the rhetoric and reality is a different matter, however. Most companies are struggling to make any progress building data analytics capabilities: 41% percent of CEOs report that they are not at all prepared to make use of new data analytic tools, and only 4 percent say that they are “to a large extent” prepared (IBM). “AI” conventionally refers to a broad class of technologies that allow a computer to perform tasks that normally require human cognition, including decision-making. Our discussion here is narrower, focusing on a sub-class of algorithms within AI that rely principally on the increased availability of data for prediction tasks. For certain, there have been major advances in the domains of pattern recognition and natural language processing (NLP) over the last several years.
Deep learning using neural networks has become increasingly common in some data-rich contexts and has brought us closer to true AI, which represents the ability of machines to mimic adaptive human decision making. Nevertheless, with respect to the management of employees, where the promise of more sophisticated decisions has been articulated loudly and often, few organizations have even entered the big data stage. Only 22 percent of firms say they have adopted analytics in human resources, and how sophisticated the analytics are in those firms is not at all clear. The promise of data analytics, by contrast, is easier to see in fields like
marketing. While there are many questions to be answered there, they tend to be distinguished by their relative clarity, such as, what predicts who will buy a product or how changes in its presentation affect its sales. Outcomes are easily measured, are often already collected electronically by the sales process, and the number of observations – sales of a particular item across the country over time, e.g. – is very large, making the application of big data techniques feasible. Although marketing is not without its ethical conundrums, the idea that companies should be trying to sell more of their products is well-accepted as is the idea that business will attempt to influence customers to buy more.
SUMMARY OF THE INVENTION
While general-purpose AI is still a long shot in any domain of human activity, the speed of progress towards specialized AI systems in health care, automobile industry, social media, advertising and marketing is considerable. Far less progress has been made in issues around the management of employees even on the first step of the AI path, which is decisions guided by algorithms. We identify four reasons why: complexity of HR phenomena, data challenges from HR operations, fairness and legal constraints, and employee reactions to AI-management. Causal reasoning is the first principle relevant to addressing these challenges across the stages of the AI Life Cycle. Because the creation of algorithms relies on association rather than causation, an absence of notions of causation makes it much more difficult to create the datasets needed for analysis: we need more data because we do not know what to choose. Causal reasoning also helps greatly with issues of fairness and explain ability. The benefits of causal reasoning do come with costs. Employers must first accept the greater costs (based on the need for more data) and lower predictive power from algorithms where we do not have causal models, and they must work to develop consensus about causal assumptions in advance of modeling. These challenges explain why the data science community is quite skeptical about causally reasoning AI systems. Randomization is a second principle that can help with
algorithmic-based decisions. First, randomizing the inputs into an algorithm is akin to experimentation and can help to establish causality. Second, randomly choosing an HR outcome with the probability predicted by an algorithm where we cannot predict outcomes with much accuracy acknowledges the inherently stochastic nature of HR outcomes and unavoidable inaccuracy of algorithms. Employees may perceive such randomization—such as flipping a coin—to produce fairer outcomes under uncertainty. Formalizing processes is also necessary to build reasonable algorithms. It ensures that the parties are aware of the assumptions built into any algorithms, the costs of building them, and the likely challenges from employees who are adversely affected by them. In the process, formalization can be enabling rather than coercive.
To what extent the changes we suggest require a restructuring of the HR function is an important question. Certainly, HR leaders need to understand and facilitate the Data Generation and Machine Learning stages of the AI Life Cycle. The integration of HR data with business and financial data should allow an HR Department to quantify in monetary terms its contribution to the company’s bottom-line. Line managers will have to refresh their skill set as well. For them, AI should imply “augmented intelligence,” an informed use of workforce analytics’ insights in decision-making. The invention on evidence-based management proposes a Bayesian approach to systematically updating managerial beliefs with new information. We consider it a helpful departure point for AI-management as well. The tension between the logic of efficiency and of appropriateness affects most organizational action. In the case of HR, the drive for efficiency and concerns about fairness do not always align. We hope that the conceptual and practical insights in this paper will move AI-management in HR forward on both counts, those of efficiency and appropriateness.
BRIEF DESCRIPTION OF THE DRAWINGS
Fig.1: depicts artificial intelligence techniques in human resource management.
Fig.2: depicts life cycle process of HR practice.
Fig.3: depicts applications pf artificial intelligence in HR.
BRIEF DESCRIPTION OF THE INVENTION
The speed with which the business rhetoric in management moved from big data (BD) to machine learning (ML) to artificial intelligence (AI) is staggering. The match between the rhetoric and reality is a different matter, however. Most companies are struggling to make any progress building data analytics capabilities: 41% percent of CEOs report that they are not at all prepared to make use of new data analytic tools, and only 4 percent say that they are “to a large extent” prepared (IBM). “AI” conventionally refers to a broad class of technologies that allow a computer to perform tasks that normally require human cognition, including decision-making. Our discussion here is narrower, focusing on a sub-class of algorithms within AI that rely principally on the increased availability of data for prediction tasks. For certain, there have been major advances in the domains of pattern recognition and natural language processing (NLP) over the last several years.
Deep learning using neural networks has become increasingly common in some data-rich contexts and has brought us closer to true AI, which represents the ability of machines to mimic adaptive human decision making. Nevertheless, with respect to the management of employees, where the promise of more sophisticated decisions has been articulated loudly and often, few organizations have even entered the big data stage. Only 22 percent of firms say they have adopted analytics in human resources, and how sophisticated the analytics are in those firms is not at all clear. The promise of data analytics, by contrast, is easier to see in fields like marketing. While there are many questions to be answered there, they tend to be distinguished
by their relative clarity, such as, what predicts who will buy a product or how changes in its presentation affect its sales. Outcomes are easily measured, are often already collected electronically by the sales process, and the number of observations – sales of a particular item across the country over time, e.g. – is very large, making the application of big data techniques feasible. Although marketing is not without its ethical conundrums, the idea that companies should be trying to sell more of their products is well-accepted as is the idea that business will attempt to influence customers to buy more.
The effective application of AI to human resources problems presents very different challenges. They range from practical to conceptual, including the fact that the nature of data science analyses when applied to people has serious conflicts with criteria societies typically see as important for making consequential decisions about individuals. Consider the following: • A first problem is the complexity of HR outcomes, such as what constitutes being a “good employee.” There are many dimensions to that construct, and measuring it with precision for most jobs is quite difficult: performance appraisal scores, the most widely-used metric, have been roundly criticized for problems of validity and reliability as well as for bias, and many employers are giving them up altogether. Any reasonably complex job is interdependent with other jobs and therefore individual performance is hard to disentangle from group performance.
The data sets in human resources tend to be quite small by the standards of data science. The number of employees that even a large company may have is trivial compared to the number of purchases their customers make, for example. Moreover, many outcomes of interest are rarely observed, such as employees fired for poor performance. Data science techniques perform poorly when predicting relatively rare outcomes. The outcomes of human resource decisions (such as who gets hired and fired) have such serious consequences for individuals and society that concerns about fairness – both procedural and distributive justice - are paramount. Elaborate legal frameworks constrain how employers must go about making those
decisions. Central to those frameworks is the concern with causation, which is typically absent from algorithm-based analyses. Employment decisions are also subject to a range of complex socio-psychological concerns that exist among employees, such as personal worth and status, perceived fairness, and contractual and relational expectations, that affect organizational outcomes as well as individual ones. As a result, being able to explain and also to justify the practices one uses is much more important than in other fields. Finally, employees are capable of gaming or adversely reacting to algorithmic based decisions. Their actions, in turn, affect organizational outcomes. To illustrate these concerns, consider the use of an algorithm to predict who to hire. As is typical in problems like these, the application of machine learning techniques would create an algorithm based on the attributes of employees and their job performance in the current workforce.
Even if we could demonstrate a causal relationship between sex and job performance, we might well not trust an algorithm that says hire more white men because job performance itself may be a biased indicator, the attributes of the current workforce may be distorted by how we hired in the past (e.g., we hired few women), and both the legal system and social norms would create substantial problems for us if we did act on it. In Amazon discovered that its algorithm for hiring had exactly this problem for exactly this reason, and the company took it down as a result. Even when the sex of applicants was not used as a criterion, attributes associated with women candidates, such as courses in “Women’s Studies” caused them to be ruled out. 4 If we instead build an algorithm on a more objective measure, such as who gets dismissed for poor performance, the number of such cases in a typical company is too small to construct an effective algorithm. Moreover, once applicants discover the content of our hiring algorithm, they are likely to respond differently in interviews and render the algorithm worthless. Most applicants already know, for example, to answer the question “what is your worst characteristic” with an attribute that is not negative, such as, “I work too hard.” Below, we
address each of these challenges separately at each stage of what we call the AI Life Cycle: Operations – Data Generation – Machine Learning – Decision Making. We rely on key ideas from Evidence-Based Management (EBM gmt) - a theory driven causal analysis of “small data”. We then suggest how, given these constraints, we might make progress in the application of machine learning tools to HR. Specifically, we focus on the role of causal models in machine learning. Establishing causation is central to concerns about fairness, which are fundamental to making decisions about employees, and machine learning-based algorithms typically struggle with that challenge. We also suggest that randomization can be useful as a decision process, given its perceived fairness and the difficulty that analytics may otherwise have in making fair and valid decisions. We base our arguments on knowledge of contemporary practice as well as on interactions with practitioners, and in particular, a workshop that brought data science faculty together with the heads of the workforce analytics function from 20 major US corporations.
Each of these operations involves administrative tasks, each affects the performance of the organization in important ways, and each includes specific offices, job roles, written instructions and guidelines to execute as well as the actual activities and interactions of all parties. These operations produce volumes of data, in the form of texts, recordings, and other artifacts. As operations move to the virtual space, many of these 6 outputs are in the form of “digital exhaust,” which is trace data on digital activities (e. g. online job applications, skills assessment) that may be used to build recruiting algorithms. Human resource information systems, applicant tracking systems, digital exhaust, and other markers are all critical inputs for the “data generation” stage. Typically, this input has to be extracted from multiple databases, converted to a common format, and joined together before analysis can take place. By “machine learning” (ML) we refer to a broad set of techniques that can adapt and learn from data to create algorithms that perform better and better at a task, typically prediction. Within
business contexts, the most common application of machine learning technologies has been “supervised” applications, in which a data scientist creates a machine learning algorithm, determines the most appropriate metric to assess its accuracy, and trains the algorithm using the training sample. Some of the most commonly used prediction algorithms, such as logistic regression and random forest infer the outcome variable of interest from statistical correlations among observed variables. The accuracy of preliminary models is assessed on the development sample until it stabilizes at some acceptable level. The final model is run on the test sample; the accuracy of the predictions on this sample is the ultimate indicator of the model’s quality. For hiring, for example, we might see which applicant characteristics have been associated with better job performance and use that to select candidates in the future. “Algorithmic management,” the practice of using algorithms to guide incentives and other tools for “nudging” platform workers and contractors in the direction of the contract, is applied to regular employees.
At present, this is principally the case in making recommendations. IBM, for example, uses algorithms to advise employees on what training make sense for them to take, based on the experiences of similar employees; the vendor Quine uses the career progression of previous employees to make recommendations to client’s employees about which career moves make sense for them. Vendors such as Benefit focus develop customized recommendations for employee benefits, much in the same way that Netflix recommends content based on consumer preferences or Amazon recommends products based on purchasing or browsing behavior. These algorithms differ in some important ways from traditional approaches used in HR. In industrial psychology, the field that historically focused the most attention on human resource decisions, research on hiring, say, would test separate explanatory hypotheses about the relationship between individual predictors and job performance. They pick the hypothesis to examine and the variables with which to examine it. This process produces lessons for hiring,
one test at a time, e.g., the relationship between personality test scores and job performance, then in another exercise, the relationship between education and job performance, and so forth. Machine learning, in contrast, generates one algorithm that makes use of many variables. The variables may not be in the cannon of the theoretical literature associated with the topic, and the researcher is not hypothesizing or indeed even examining the relationship between any one variable and the outcome being predicted. Indeed, one of the attractions of ML is its investigation of non-traditional factors because the goal is to build a better prediction rather than advancing the theory of the field in which the researcher is based by providing evidence on particular hypotheses. “Decision-making,” the final stage, deals with the way in which we use insights from the machine learning model in everyday operations. In the area of human resource decisions, individual managers may have more discretion now in how they use empirical evidence from data science and other models than they did in the heyday of the great corporations when hiring and other practices were standardized across an entire company. Managers today typically have the option of ignoring evidence about predictions, using it as they see fit, and generating their own data about actions like hiring in the form of interviews they structure themselves.
, Claims:We Claim:
1. AI enables the collection and analysis of data in your HR processes to eliminate biases and guesswork to guarantee you are choosing the right candidate or offering the best compensation and benefits plan.
2. For example, mining recruitment data helps uncover challenges so you can address them objectively.
3. AI techniques are those procedures which are used to enable computers to show human like intelligent activities such as visual perception, speech recognition, decision-making, natural language understanding etc. 4. AI recruiting is the process of using artificial intelligence to automate time-consuming, repetitive tasks while offering personalization and data insights throughout the hiring process.
5. AI can help the MIS to provide decision support by analyzing and processing large amounts of data in real-time, and presenting insights that can be used to make better business decisions.
6. For instance, let's say a retail company is using an MIS to track sales data from various stores.
7. Artificial Intelligence can help the HR team design and automate their onboarding process to be more organized, efficient, and personalized. There are a lot of steps in onboarding that can be automated and handled by an AI system. Some of them include: Verifying documents.
Dated this the 12th June 2023.
Senthil Kumar B
Agent for the applicant
IN/PA-1549