How will AI change HR in the future?

In recent years, artificial intelligence (AI) has transcended its status as a mere technological innovation, becoming an integral part of our daily lives and work processes. Its impact on the field of human resources (HR) is not just expanding but also promising a transformative future. In this context, Genesis Lab's Geunsik Yuk, General Manager of HR Business, and Min Kyung Yang, CEO of HRBletin, delve into "How AI will revolutionize HR in the future."

--- ---.

Why did you request the interview?

Yang Min Kyung (CEO, HRBlattin)

First, I would like to explain the background of my interview request. Although AI has been a hot topic for about five years, and many companies have been using it as a marketing tool, it hasn't been long since I realized, "This is AI." In particular, a significant change was coming when things like YouTube's algorithm became known to the general public and technologies like ChatGPT started to emerge. The HR tech space is also evolving, and I'd like to ask you two questions: first, about HR tech trends, and second, about how you're applying these technologies to create value from ViewerHR's perspective.

👨‍🦰육근식 Director

Okay, so your first question concerns HR tech trends, specifically the flow of how AI is being applied in the HR space.

Yes, I'm talking about the trend of applying AI in the HR space and trying to do something new.

👨‍🦰육근식 Director

Let's talk about ViewerHR, the HR AI we are aiming for. ViewerHR is an HR AI total solution. We see a lot of work in HR where people interview, assess, and diagnose other people. For example, you might have an exit interview; when you move into recruiting, it's an interview, and when you're dealing with employees, it's an employee performance review, competency assessment, or interview.

The business model is that AI helps with these things. About 70% of AI is used in recruiting, and the other 20-30% expands into other HR areas outside of recruiting. The primary mechanism is that AI helps humans assess, diagnose, and interview other humans.

Yeah. So we're talking about face-to-face. You can do written assessments, you can do written interviews, you can do those things, but the difference with ViewerHR is that we're doing it through video, so it's like a human talking to a human, right?

👨‍🦰육근식 Director.

Exactly, and that's what AI does for you and what we're doing with that concept.

---

What makes ViewinterHR different

Yes. Then I'm curious: When ViewinterHR came out, what point were customers particularly surprised by, or what was the differentiator? What's the main thing that customers are saying, "Oh, this works, this is good to do this."

👨‍🦰육근식 Director.

We still need to be 100% like a perfect human assessor or interviewer - it's recorded chiefly or taped, so it only partially replicates the nature of a face-to-face interview where you can take the answers and ask additional questions in real time. However, recent advances in generative AI technology will likely address this shortly. These technology trends have given us a clear direction for our technology. Our technology employs a deep learning approach where the AI learns insights from upscaled interviewers. In this process, hundreds of thousands of candidate videos and answers are evaluated and labelled by experts, which the AI learns to evaluate candidate videos and answers.

Invited people who are good interviewers in academia and practice and trained the AI with their data.

This will solve the problems of inconsistent evaluation and interviewers' lack of objectivity because the data we trained on already belongs to good people.

👨‍🦰육근식 Mr. Lee

Another aspect that sets us apart is our commitment to transparency. We use a deep learning approach, learning from the best interviewers, and we make this algorithm as transparent as possible. By sharing our process, we aim to build trust with our customers. We believe in openness, unlike many other AI services that keep their processes hidden.

Publishing it makes it trustworthy, but isn't it possible for someone else to try to learn something similar using that algorithm?

👨‍🦰육근식 Director

Collecting tons of video data and recruiting experts is a significant barrier to entry. We are not a simple rule-based AI, so we put much effort into objectivity and removing bias through this process. This makes our technology different, and because it's deep learning, it can't be easily replicated even if we disclose it.

So, did you compare your solution's results with the interviewer's evaluation before deciding to use it to demonstrate its reliability?

👨‍🦰육근식 Director.

Yes, we test that in the pilot or proof of concept phase. We also validate various assessment tools, including PeopleAnalytics, and have received reliability certifications from national organizations. This validation process ensures that our clients can trust our AI's assessments.

---

Regarding AI and HR trends

What can you tell us about AI and HR trends?

👨‍🦰육근식 Mr. Director

Digital transformation has been a buzzword, and we're witnessing significant shifts in HR. Previously, HR management was primarily collective, but now it needs to be personalized, necessitating systems and analytics. While most departments are data-driven, HR has often relied on experience for decision-making. However, integrating systems, data, and AI is becoming crucial. Many companies have systems in place but need the ability to analyze that data. This is where AI can play a transformative role, and I want to underscore the need for HR tech solutions in this context.

If you look overseas, there are tons of tech solutions across all areas of HR, especially in the US, where many HR tech startups have emerged in recent years. In Korea, we have traditionally favoured building our own, but with the rise of startups, we are seeing an increase in renting tech solutions as SaaS. This has been a trend in the last 3-4 years. Some companies offer software only, while others have AI and various solutions depending on the situation.

The bottom line is that how you utilize these solutions will determine the success and productivity of your HR digital transformation. You can develop your own or rent specialized solutions for each area, the latter of which can be more efficient. According to BCG Consulting, the more sophisticated regions of HR become, the higher the impact and the less accessible they are.

---.

Future expansion plans

I've noticed a shift in the market, and I've recently heard that many companies that used to be called tech companies will now be called data companies. Data will be the key to providing differentiated value; how much data you have and how well you utilize it will determine your competitiveness. I understand that ViewinterHR is providing great value based on data. You are also working on measuring the effectiveness of interview learning and improving performance evaluation capabilities. What are your future expansion plans? How do you plan to leverage this technology?

👨‍🦰육근식 Mr. Director

We've always been clear about the direction of our business model, but recent advances in productivity AI have accelerated our plans. We aim to leverage AI to evolve into a more sophisticated assessment solution. With our diagnostic-based ACDC (all-inclusive system for assessment, development, and promotion) approach, we're moving towards AI-assisted structured assessment tools and real-time interview possibilities. We see this shift happening even faster thanks to advances in generative AI. There are very few companies that combine AI and HR, so our move in this space will be a pretty differentiated approach.

So, the core technology is assessment, and you're planning to assess higher-order competencies, from recruitment to performance appraisal to employee development. As you accumulate data throughout the employee's lifecycle, from interview to exit, you can analyze trends and changes in an individual's career development and job growth. Is this the direction you're going?

👨‍🦰육근식 Director.

It's totally in line with our vision. The amount of data means that many people are using the platform, and utilizing this data effectively is only possible with AI. Our solution viewer for interview practice already has over 500,000 cumulative users and 3 to 4 million interview videos they've created. This means we're expanding beyond the hiring scene into post-employment data collection. ViewerHR started with Viewer initially.

---

The difference between Viewer and ViewerHR

So it was for candidates. What's the difference between Viewer and ViewerHR?

👨‍🦰육근식 Director

Yes, Viewer is a practice site or practice solution for candidates to practice face-to-face interviews. It starts accumulating data before they're hired and continues to accumulate data throughout the hiring process and after they're hired. What's even more interesting is that this practice program is also being rolled out to colleges and universities, so we're also accumulating data from college students.

Are they assessing job competencies, or are they mostly looking at general, common competencies?

👨‍🦰육근식 Director

It extends to the workplace and is even being utilized in high schools. Currently, solutions are enterprise-centric because companies need to select talent, but that's changing as individuals empower this data. Rather than simply representing your skills with a resume or GPA, it makes more sense to assess yourself through objective AI assessments constantly. This accumulation of data paves the way for individuals to take control of their life cycle.

---.

A summary of the differences between Viewer and ViewerHR:

Viewer: A site for practising in-person interviews.

ViewerHR: AI interviewing solution used by companies to hire

---

How will this change the job market?

What are the changes for job seekers? Can you explain more about the changes for job seekers? If companies start using ViewerHR more, how will that affect job seekers?

👨‍🦰육근식 Director

Job seekers are curious about the myths and truths around how AI interviews work today. In many cases, their understanding is split down the middle. Let's look at some misconceptions job seekers have about AI interviewing.

• ⓐ Myths and truths: Many job seekers need more information to understand AI interviews. We try to clear up these misconceptions through the media and platforms like Job Planet. For example, there is a common misconception among job seekers that AI analyzes emotion, which is different. We clarify that AI interviews are not based on emotion and that not all answers need to be emotionless.
• ⓑ Difference from in-person interviews: Many candidates feel that they can get a "feel" for whether or not they'll be hired after an in-person interview, but with AI interviews, there is no such interaction, which makes them feel anxious. We're trying to alleviate this anxiety through technology.
• ⓒ Unpleasant in-person interview experience: A surprisingly large number of candidates have had unpleasant experiences. In contrast, AI interviews give every candidate the same opportunity, which is why many job seekers prefer AI interviews.
• ⓓ Expanded opportunities: Many companies need to screen applicants during the hiring process, and AI interviews or personality tests give more candidates a chance to showcase their skills. It also expands opportunities for job seekers who might have yet to get a face-to-face interview.

---

Criteria for choosing a good AI

Like personality tests, self-reported assessments vary widely in price, but ultimately, it's all about validity and expertise. So, what are your criteria for choosing a good AI company?

👨‍🦰육근식 Director

There are several critical criteria for choosing a good AI company.

• ⓐ Reliability and validation: First, it's essential to ensure the AI is reliable, well-built, and highly valid. Customers shouldn't just take the company's word for it but prefer AIs that have been double-verified by independent organizations such as national agencies. For example, we have been verified and awarded by the TTA and the Ministry of Science and ICT to ensure our credibility.
• ⓑ Validity and relationship with selection tools: Secondly, you should look at how valid the AI is with other selection tools and whether the data has been adequately validated. Many HR AI vendors tend to overestimate the power of their systems, but they often need independent data analysis to back up their claims. We analyze our client's data to ensure high performance and predictive validity and provide results.
• ⓒ Customer-specific analysis: A good AI company should not simply claim high validity based on its data but should analyze the validity in real-world usage situations for each customer and provide results. Since each company's situation is different, there is no one-size-fits-all AI solution, so it is essential to have an approach that validates validity through real-world usage results.

When choosing an AI vendor based on these criteria, you can select a vendor that offers reliable and validated AI solutions.

---

renewal 

If you can share, what is the renewal rate? If it is a good service, I will continue to use it, but is the renewal rate high?

👨‍🦰육근식 Mr. Lee

Our retention rate is almost 95%+. The downside is that the decision-making process for adopting these tools or changing to another tool is quite long. We carefully consider them for as little as three months or as long as a year, and during this process, we run a proof of concept for actual use. Therefore, once adopted, they tend to be used for a long time, which keeps the retention rate high.---

---

When AI is used in face-to-face interviews, how do interviewers react? 

How do interviewers utilize the data provided by AI when AI and the second interview conducted the first interview is performed by a human?

👨‍🦰육근식 Director

Interviewers have been very positive about using AI in face-to-face interviews. Compared to other tools, there's a big difference in how usable and acceptable AI is. For example, personality or gaming neuroscience-based aptitude tests have a low acceptance rate because people must learn how to use them in practice. However, our AI uses a BEI (Behavioral Event Interview) method, which means that all questions and answers are recorded on video and can be viewed by the interviewer. This makes it easy for interviewers to understand the results of the AI's assessment and use them in their evaluations.

In addition, our AI can be customized to each company's talent profile. The AI can conduct a preliminary interview to assess whether the candidate fits the company's talent profile and ask questions that the AI cannot cover in the face-to-face interview to provide a comprehensive assessment of the candidate. Interviewers are very optimistic about the use of AI in this regard, and because it is a BEI method, there is a high level of acceptance of the AI-based interview method.

---

Hard-skill AI assessment with generative AI

I would like to know more about hard-skill AI assessment with generative AI and how it can be utilized in the enterprise.

👨‍🦰육근식 Mr. Director

Another feature of our solution is that managers can create their questions. In general, many tools provide personality tests or common competency assessments. Still, it's difficult for HR teams to develop questions about job competencies because they need expertise in each job. To meet this need, we created a solution that leverages generative AI to automatically generate job-related questions, evaluation criteria, and model answers. This allows them to conduct job interviews more effectively.

How does a company manage this system? For example, if a manager wants to create job questions for a specific department, can they manipulate the system, or do they have to request it?

👨‍🦰육근식 Director

Managers can create questions in the system themselves. Once you provide the system with the necessary inputs, the software will automatically generate and evaluate the queries. This process is done within the software so each department or team can manage it as needed.

---

Trends in the AI Industry

👨‍🦰육근식 Mr. Director

Speaking of trends in the AI industry, the emergence of generative AI is rapidly creating new business models. The AI industry, which took a break after the AlphaGo boom, has been revitalized, especially with the recent emergence of ChatGPT. However, only some AI services have a tangible impact on the masses. For example, recommendation algorithms on Netflix and YouTube or personalized product recommendations in online e-commerce have been around for a while.

In the future, standalone AI services will drive the industry. If you think of the AI industry as a pyramid, at the bottom are the foundational sectors like semiconductors, at the top are mid-levels like Large Language Models (LLM), and at the top are companies that provide tangible services to society. In this structure, the development of the lower-level industries is essential to the growth of the higher-level industries. In this context, companies like Microsoft Coria are consulting with companies like us and making their services freely available, hoping that the growth of our business will lead to the development of theirs.

We'll see more services that go beyond recommendation algorithms and help users. While some functionalities, such as AI announcers, are being developed, much work remains regarding how they connect to services and support users. We will see more independent AI services like ours in the future.

---

Last comment

AI, especially generative AI, has made a big impact. In the HR industry, there were expectations such as improved matching rates, but the direct impact on revenue was more about the difference from other comparators. Apart from the fantastic experiences AI can deliver, I think only some things touted as AI were a wow factor. People are feeling this shift. Is there anything else you'd like to add?

👨‍🦰육근식 Director

Excel and PowerPoint used to be critical tools for the productivity and performance of tech workers in Korea, but now AI is a must. And we're not just talking about AI; we're talking about data and systems, including the data and systems connected to AI, which we see as essential. There will come a time when you can only do your job using AI. We need to stop racing the racehorse called AI and get on the back of the racehorse called AI.

Yes, stop racing the racehorse called AI and get on the back of the racehorse called AI. What do you think, Mr. Choi?

🧑Leo Choi

As I dealt with AI and HR, I realized that hard skills can be replaced by AI. What is important is strategic thinking and prioritization, leaving hard skills to AI and focusing on important things other than hard skills. This is the way to maximize business performance.

👨‍🦰육근식 Director

It's essential to utilize employee information, such as HR data, effectively. With tools like Copilot, AI can analyze your vast amount of data and help you make better HR strategies and decisions. AI will eventually be utilized to make decisions faster and more accurately.

The advancement of AI technology and its application in HR rapidly progresses, creating many opportunities for organizations. However, there is also a need to address social acceptance and the unique human dimension of AI adoption.

---

Interviewee Yuk Geunsik (Head of HR Business, Genesis Labs)

Interviewer Yang Min-Kyung (CEO of HRBletin)

Review & Edit Leo (Genesis Labs Marketing)