Note: This article is a pre-print of my latest paper, the pre-print carries DOI: 10.13140/RG.2.2.15401.51045. Here the article is presented as a preprint with rights as above.

Abstract- In this article, we provide ways in which AI and Robotics can help in the tedious work of interviewing in ethical ways. The interviews can be about employment, performance, or even job termination. The article discusses many use cases and advantages for handling the solutions for the problem in ethical ways. Ultimately, we affirm ethical concerns that Robots and AI should always concede to the human guide and in no way try to take over the human guide. This means that Robotics and AI should work under guidelines of ethical contribution to the human world. Robots should be made with such considerations. This article is motivated by Elon Musk’s current abrupt job termination notice to Federal Government employees.

1. Introduction

Interviews can be of various kinds. There are job interviews, when a person applies for a vacancy in a company/office, performance interviews, to compute bonus, annual meeting interview, to compute yearly progress, even to exit interviews, when a person decides to leave a company or when the company decides to let go of an employee or contract worker. The interviews are held not just of the employees but also of contract workers and students as well. Now, the human touch is always missing if we think of a robot conducting an interview. Another thing is the superiority complex present in robots due to the immense knowledge they possess; this may and can make the person in front of it a bit uncomfortable, hence the need for emotional intelligence in robots to be inculcated.

The aim of this process is warmth in dealing with humanly things such as exit interviews, performance falls, and so on. Given that robots are well aware of what it means by poor performance, the sentiments that arise in it are vivid in their minds. The time it hears words poor performance or say a gap year, all kinds of negative sentiments would arise in Robots Mind, hence special algorithm need to be written to clear this neuron firing and to give it a task to focus on interview process.

Next, what we can think of is “the ethical issues behind these processes of human interviews to a robot”. The ethical issues are immense and need to be considered. First, the methodology needs to be mentioned well in advance to a robot, who should be well trained on the processes to be followed for different kinds of users it shall serve. Yes, it’s a two-person one-to-one relationship when one robot interviews one person or a one-to-many relationship when, say, an exit interview is performed from one robot to many people. Typically, these interviews don’t need to be many-to-many, as one robot can topple up the work of many robots in these circumstances. The many-to-many case may arise in case several multi-disciplinary robots interview many candidates catering different views, say, in a group discussion part of a Business Management Degree Entrance Exams, and so on.

The processes conducted by a robot with or without the guidance of a human are referred to as robotic processes in this article. The key areas of robotic process interviews and their scope in ethical interviews are the following:

1. Robots for job/admission interviews: One or more robots take job interviews or admission interviews in some schools or universities. These interviews determine the future of the candidate and, at the same time, determine the future of the company and the company in which they came to work. Their creation as a part of the work would grow the company to new heights. Hence, we need the right Robotic AI to handle the world’s future, one company at a time.

2. Robots taking appraisal interviews: These are like annual appraisals or annual performance interviews, wherein grades/bonuses/renewals of tender are discussed. These are needed to boost the employees for productive work, to encourage employees to work with more sprits, which Robots lack.

3. Robots taking exit interviews: When a candidate leaves an office, it’s not just a resignation; it’s a resignation from the bonds of work, togetherness, and processes that were followed. It’s not a one-day goodbye, and Robotic AI can help the Exit process as a warm Thank You for the togetherness in the job, whether it’s regular, part-time, or remote.

The approaches of the three AI processes are different and need different strategies. Let us study the ethical interviewing processes and understand how to implement the ethical perspectives.

Now, why all this when humans can do so much more flexibly? The thing is, times have changed, and hence the way we do work, knowledge is increasing daily, the amount of information is changing with AI progressing, it’s difficult for humans to keep abreast with the latest tech/news/world issues and more. Hence, AI-based experts can help in selecting the best candidate, say, to go to Mars to stay or to be allowed to go to the Moon for a picnic as a part of the appraisal process.

In the following sections, we will discuss the ethical considerations, challenges in such implementations, use cases, and future work.

2. Ethical Consideration

Here, we provide ways in which ethical considerations for the above-defined problems are considered. AI and robots must perform under the ethical guidelines laid out in a place. That should be in no way AI and Robotics can let aside laws to generate its supremacy. AI should not be allowed to become the Boss of all organizations. This is a big threat to humanity by AI, and hence, we must follow in and lay bricks of ethical requirements as building blocks of the tower of ultra-modernization.

The interview process between one or more robots and candidates needs to follow the following guidelines:

2.1. Robots taking job/admission interviews

The interview process must understand the human candidates and humanely talk to humans to access the human and student’s hidden potential as per the interview demands. The robot who takes interviews of any kind mentioned above should be free from any kind of bias, be it gender based bias or medical condition-based bias. The only thing the interview owner robots need is performance.

The data learned by a robot in the interview process is subject to privacy and must be deleted from the robot’s memory as soon as the interview is over. Maybe in some cases, the data learned can be converted to synthetic data by changing the name and identity and using just the conversation to generate data to improve the entire interview process. Security of such an interview data is of utmost importance.

2.2. Robots taking exit interviews

Consider a group of people being laid off by a company they have worked for years. Is it ethical to notify them via just an email? No, right? Furthermore, there is no one on the team or in the office to speak personally with those being laid off. Imagine if HR lacks sufficient human resources. So, can HR seek assistance from robotics? Yes, if they have it available. Why? Firing without such a protocol is not easy in any sense. Therefore, we can establish one of the three protocols:

1. Ensure the exit process is as smooth as possible, even if the layoff was due to poor performance; he is human and should be treated compassionately.

2. Robotics can effectively assist HR with the exit interview process. The only concern is if the employee is upset about being fired, in which case it’s important to ensure the robot is properly managed. In this scenario, a human might express satirical or harmful comments about the firing process, so teaching the robot how to handle these situations is essential.

2.3. Robots taking appraisal interviews

AI-based robotics can make appraisals and academic/non-academic performance report cards more manageable and thorough. The robot can review entire performance records of employees and students, catering to the needs of the appraisal manager or academic head. In this way, the robot can provide strict feedback on task performance. Given that robotic intelligence cannot be challenged, the robot may seem superior to the candidates if it can access vast information. However, robots may not possess true intellect. Therefore, from an ethical standpoint, this process must be approached with humanity when evaluating performance. Training is essential to ensure that robots remain humble and kind to the employees whose appraisals they assess. It is important to recognize that a person may receive a poor appraisal due to personal circumstances or might excel in a different area.

3. Creation of such Robots

Robots that perform the various tasks mentioned above must do so ethically. Therefore, embedding ethical considerations deeply within robotic systems is essential. To achieve this, guidelines must be established for this ethical domain. Reinforcement learning can be a beneficial approach for robots involved in any of the aforementioned types of interviews, including job or admission interviews, appraisal interviews (such as discussions about report cards), and exit interviews for a warm farewell.

Each of these interviews has its unique requirements and can be viewed as a specialization of an interview robot. Each robot can switch to specialization mode to enhance the interview process. Alternatively, separate robots can be maintained for each specialization, depending on what is most convenient for the head of the organization (university).

There should be a checklist of do’s and don’ts for ethical processes that the HR Robot must follow for each type of interview. There are several factors to consider if only one robot is responsible for all kinds of HR tasks, including computation, addressing the individual being assessed for performance, and determining the processes to follow when the human manager or owner wants to know the robot’s decision in addition to their own.

Before we delve deeper, let’s discuss some fundamentals of reinforcement learning in relation to interviews and robotic ethics. Reinforcement learning provides the learning system with continuous feedback based on the output produced by the system. This feedback is designed to improve the system. In the context of a job interview, it can be understood as follows:

1. A person meets a robot, and a human expert analyzes the robot’s reactions. Then, feedback on these reactions is sent back to the robot, creating a feedback loop. This process can improve how job interviews are conducted.

2. For a robot conducting interviews, the questions should relate to the tasks performed by the human worker or student. Another ethical concern is that the robot must listen to the candidate’s words, comprehend them, and evaluate their performance. It is unethical to note the grade simply; the entire response should be recorded for reinforcement learning when human experts in the field analyze it. This would convert the robotic interviewer into an expert in handling termination interviews.

3. Robot accessing the score. If the interview is not conducted, only results from yearly or half-yearly progress will be provided by a human expert. Then, the robot will learn the score of the mentioned performance. The interaction should be kind and warm, informing the individual that they will always be alumni of the company and mentioning a few perks and support they will receive, whether one-time, yearly, or for life time.

4. A robot must provide a warm and humane response to candidates regardless of whether the feedback is positive or negative. The robot must approach this sensitive issue with empathy, as candidates often feel distressed upon learning of their termination.

These are some examples of reinforcement learning in AI and robotics. Many more examples and use cases can be provided; we believe the examples above are sufficient to convey an understanding of ethical solutions in AI and interviewing. In the next section, we will examine some limitations of AI and robotics in interviewing both human and robotic candidates.

4. Limitations of AI and Robotics

It must be programmed into the algorithm of robotic interviewers that they are robots and that they must follow the instructions of the admin or guide xyz. Any message from the admin or guide xyz should be viewed as an order, not an answer. Here, xyz can refer to a group of people who can command an interview robot, or it can be an owner who may or may not be part of the admin. Therefore, a robot must be made aware of two states of humans, as follows:

1. The humans who create robots and AI are called the owners. These robots are registered under the name of the admin, or sometimes referred to as xyz. “Admin” is a more familiar term for software developers, so the robots must follow the orders of admin/xyz.

2. The other group consists of people who encounter the AI-based robot. These individuals can be interviewees, staff in an office, institute, or school where the interview is taking place. The robot needs to communicate with them according to the job role and engage with the candidates based on the performance requirements.

The limitations of AI and robotics depend on the authority of the administrator or owner of the AI-based robot. The robot can be programmed to stop asking questions in training mode and can learn to lower its voice in response to conflicting issues that may arise during human conversations.

These analogies stem from the code of ethical AI use. The purpose of AI is to assist humanity by serving in the best interest for which robotics and AI were developed. Thus, a robot can learn in a supervised manner and later in a semi-supervised manner for the process of interviewing and conversing with humans. It is important to note that the robot (or AI) can pause and switch to the next topic as needed.

The problem here lies in determining when to enter admin mode. When should the admin take control of the robot? When is it appropriate to guide the robots remotely? When should the robot act as an extension of a human admin? All these considerations are necessary since robotics has not reached that level yet. These limitations emphasize that entering admin mode or being guided by an admin is currently the only viable solution until robotics matures as a field. The field of robotics has grown right before our eyes, and it’s still in its early stages. Once these processes begin, it will take time to reap the benefits. Therefore, we must view these limitations as opportunities and a step forward in robotics. We should focus on overcoming these limitations to establish a solid foundation for humanity’s expansion across the Universe.

Conclusion

AI and robots should be designed to accept human guidance. However, a robot can take on a leadership role during interviews with people. Engineers have several ways to embed ethical considerations into robotics algorithms. We have also explored various methods to prevent robots from causing harm to humans. Similarly, we can view this process as just the beginning of other similar automation efforts. All of this aims to ensure that humans and robots can effectively and peacefully perform essential tasks for the welfare and management of Earth. The ethics of interviewing are important, and we have discussed different types of interviews. We conclude that in any kind of interview, humans should direct robots to ensure ethical interviewing, even if this requires humans to take control of robots and operate them remotely for a period of time or longer.