gift_box
en

Advantages, disadvantages, and prospects of using ChatGPT in recruiting - Volodymyr Dybenko

Volodymyr Dybenko, a co-founder of Artel Outstaffing Global, TalentSearch, and Salee, told LC Work about the tasks recruiters can delegate to artificial intelligence, testing СhatGPT results, capabilities, prospects, and disadvantages of using this tool in recruiting.

Advantages, disadvantages, and prospects of using ChatGPT in recruiting - Volodymyr Dybenko

Valentyn Ihnatiev

17.03.2023 | 12:59

4427

0

With the proliferation of artificial intelligence and machine learning, recruiters are turning to innovative methods for finding and selecting candidates. One of the most effective tools that helps automate the recruitment process is ChatGPT and its derivatives.

ChatGPT is an interactive chatbot system that uses OpenAI artificial intelligence to interact with users. This means that recruiters can use ChatGPT to automate many routine tasks related to the process of finding, attracting, and selecting candidates.

Read also: What is ChatGPT and what are its functions?

One of the most obvious benefits of using ChatGPT is increased efficiency of outreach. Thanks to artificial intelligence, ChatGPT can quickly write personalized messages for both email and LinkedIn, allowing recruiters to focus on the most promising candidates.

Finally, ChatGPT can be used to improve the quality of candidate selection. Thanks to artificial intelligence, ChatGPT can analyze not only the information provided by candidates but also their LinkedIn profiles. This helps recruiters better understand candidates and their potential in the workplace.

All in all, ChatGPT is a powerful tool for recruiters to automate the hiring process and increase the efficiency of candidate selection.

Tell us about your professional career: where did you start, what companies did you work for, and what area of recruiting do you specialize in now?

My professional career in IT started in 2001 when I launched my first startup.

Since 2003, I have been involved in outsourcing both to large companies and creating my own companies. In 2013, I opened my outstaffing company AOG.jobs, which has transformed into Artel Outstaffing Global. As you may know, recruitment is the key process in outstaffing. Since then, we have been focusing on recruitment automation, and as a result, in 2017, we launched TalentSearch.pro (formerly TalentScan.pro), a full-cycle recruitment platform.

In 2022, we focused more on automating candidate outreach, as we noticed that recruiters were sending template-like messages. And while we were operating in the Ukrainian market, it was still working, but when we entered the international job market, it became a problem.

Which of their daily tasks can recruiters facilitate with ChatGPT?

Recruiters can use ChatGPT to:

  • Generate personalized messages for candidates that contain information about vacancies and job requirements;
  • Write job postings;
  • Analyze candidate profiles on social media and other sources to understand more about applicants and their potential in the workplace;
  • Match vacancies to candidates;
  • Create a search plan for a specific candidate;
  • Find answers to the question of where else to find a candidate.

This helps recruiters focus on more important tasks and attract more candidates that meet the job requirements.

What tasks ChatGPT can't handle?

In general, ChatGPT can handle most text-related tasks. At the same time, ChatGPT may have problems evaluating job candidates because it has limited access to candidate data, such as information about their activities. Typically, ChatGPT works with data collected by the end of 2021, and this information may be outdated or incomplete.

For example, if a recruiter wants to evaluate a candidate with ChatGPT, the system may not have access to the latest changes to the candidate's LinkedIn profile or other resources. This can lead to an inaccurate assessment of the candidate and force the recruiter to use other sources to obtain up-to-date information.

Thus, ChatGPT can be a useful tool for recruiters, but it has its limitations that must be taken into consideration.

For example. First, I asked about the candidate's soft skills, and here it (ChatGPT) did an excellent job. I took this profile as an example.

ChatGPT

But then I asked chat if this candidate would be suitable for a front-end developer, knowing well that he was not.

ChatGPT

I did not find such skills and asked the Chat again.

ChatGPT

And ChatGPT apologized and said that the candidate really doesn’t have such skills.

On the one hand, if we copy all the profile information and make explicit markup, ChatGPT will do a great job with the analysis. On the other hand, if we provide a complete job description, it will do the correct matching.

Read also: Business emails: how to write to be answered.

What can you say about the practical results of ChatGPT testing? What tasks did the AI do best? And on what tasks did it perform the worst?

In practice, this tool inspires and helps consider different options for solving tasks. For example, if you need to write an article and don't know where to start, it can suggest different approaches to covering information based on the initial data. Thanks to this, you can broaden your vision of the topic and find new insights that help deepen your knowledge and understanding. You can also use this tool to get new ideas and additions to existing materials, which makes it indispensable when working on any project.

What are the shortcomings of ChatGPT at this stage?

At this stage of ChatGPT's work, there are several drawbacks:

  • insufficient accuracy of answers in some situations, which can lead to incorrect decisions in the candidate selection process;
  • limited capabilities in generating complex scripts, such as X-Ray requests, which are required for candidate selection in some industries;
  • lack of ability to adapt to the language of candidates from different countries and cultures, which can lead to misunderstanding;
  • insufficient work with data, which can lead to inaccurate selection of candidates and incorrect handling of requests;
  • if you use ChatGPT as a chat and not as a neural network from OpenAI, you need to copy and transfer a lot of data to the query.

These disadvantages should be considered when using ChatGPT in the work of a recruiter, and other tools should be used to complement its capabilities.

How to properly form a request to ChatGPT to get the best possible result?

To get the best possible result from ChatGPT, it is essential to formulate your request clearly and concisely.

It's best to start your request with a specific question or topic to give the system a clear understanding of what you're looking for. You can also use keywords and phrases to help the system understand the context of the question. It's also important to include as many details as possible about what you're looking for in your question, as this will help the system understand your needs and provide relevant answers.

For example, instead of asking: "Write me an invitation for the candidate to join me for a networking event," it is better to formulate the question like this: "Write me an invitation for the candidate to join my network. My title: Recruiter at Artel Outstaffing Global. And I want our vacancies to help candidates achieve their goals." The more specific you are, the more accurate the answers will be.

Give us a successful example of working with ChatGPT: creating a request and receiving a response for a job posting, writing an invitation to a candidate for a specific job, etc.

Let's start with an example of writing the first message to a candidate.

We provide parameters about the candidate and a description of our vacancy. In fact, you can provide fewer details. It all depends on the outcome you want to get.

ChatGPT

As a result, we get a response:

ChatGPT

We understand that this is too much and ask you to shorten the text so that our candidate can read it:

ChatGPT

You should also check translation into other languages:

ChatGPT

The next example is writing an invitation to a networking event.

Since we have a limited number of characters for the invitation, we should immediately notify ChatGPT about this:

ChatGPT

I don't like this invitation's tone, so I will give it a revision.

ChatGPT

And this option has already satisfied me.

Give us an unsuccessful example of working with ChatGPT.

We tested whether ChatGPT could generate an X-ray request based on a job posting, and it didn't work.

Request to ChatGPT:  Could you please make an X-Ray script for these parameters: "Python Developer (Senior); 3+ years of experience with Python; strong understanding of OOP; understand asynchronous programming (async/await); experience developing server-side, preferably microservice architecture; experience with at least one of Tornado, AIOHTTP, Django, Flask frameworks; familiar with databases such as Postgres, Cassandra, MongoDB; familiar with ORM; understanding of third-party APIs; understanding of data formats XML/JSON; familiar with Linux OS".

ChatGPT's response:  I'm sorry, as an AI language model, I don't have the capability of creating X-Ray scripts. However, I can provide you with a sample script that would help you to filter resumes on the given parameters:

ChatGPT

I provided an example when ChatGPT said that a candidate had programming skills, but in fact, they did not. You have to be careful and consider this when working with AI.

Have you tested other AI programs to perform work tasks? If so, which ones and for which tasks?

Actually, there are a lot of tools; most of them use the neural network from OpenAI, which is used in ChatGPT, and every day 10 new tools appear. I track such programs on Futurepedia. You can do it by the corresponding code.

I personally really like ComposeAI for text generation. It allows you to create text anywhere with the help of a neural network.

We created an application ourselves to generate invites, first messages, and follow-ups on LinkedIn. This application is currently more focused on salespeople, but in the new version, you can already choose your role, including the role of a recruiter. The advantage of this app is that you don't need to copy all the information from the candidate's profile in ChatGPT because Salee automatically analyzes who you are writing to and generates a message that impresses you with its personalization in one click.

Do you believe that AI will lead to a gradual decrease in the number of recruiters in demand? Why?

The tools already available on the market increase the efficiency of recruiters in some processes by 50%. This means that one recruiter can process more candidates. And it's hard to imagine how many more tools will be on the market in the future. It is only natural that there will be tools that will increase the efficiency of recruiters in the company, and those that do not use AI will probably be less effective.

I also believe that AI can lead to a decrease in the number of recruiters in demand but it will not put a halt to their work. Recruiters will still be needed for more complex tasks like interviewing and evaluating candidates. In addition, AI may lead to the emergence of new roles in recruiting, such as data scientists, who will be responsible for analyzing data and using AI to improve recruiting processes.

Of course, AI is not a one-size-fits-all solution for all companies, and its use requires specific knowledge and skills from recruiters. Also, the possibility of using AI depends on the type of vacancy and the candidates looking for a job. So, recruiters must understand when to use AI and when not, to ensure efficient and successful work.

In your opinion, can ChatGPT really become an integral tool in the work of a recruiter? Or will the loud discussions soon die down, the hype will subside, and recruiters will return to their usual routine without using AI?

I believe ChatGPT can become an integral tool in the work of a recruiter, as it can help focus on more critical tasks and attract more candidates who meet the requirements of the vacancy. However, at the moment, it has its limitations that need to be taken into account, and recruiters need to understand when it is appropriate to use ChatGPT and when not.

As for whether the loud discussions will soon subside and recruiters will return to their usual routine without the use of AI, it depends on many factors, such as the availability of technologies that can replace AI and the readiness of recruiters to use these technologies. Nevertheless, the role of AI in recruiting is currently growing, and it is likely that in the future, AI will become an even more popular tool among recruiters.

It can be ChatGPT or another neural network, but new tools for recruiters should use AI. Only such tools will be interesting for recruiters.

Rate this post:

0

Comments