With recent hype around ChatGPT, the artificial intelligence (AI) chatbot that can provide human-like responses to queries and generate content, many employers and talent leaders have questions and concerns about how AI can help or hurt recruitment. For example, how much time can you save in creating customized job descriptions? Can candidates get hired by using AI-generated cover letters? Could these tools replace some knowledge workers altogether?

As AI becomes increasingly in demand in the tech space, Indeed recently appointed Data Science Director Trey Causey in the new role of head of AI ethics. Based on his research and experience, here’s a straightforward guide for what AI is, and isn’t, capable of accomplishing in recruiting and hiring.

AI Can’t Fact-Check Information

ChatGPT works by sorting through the vast amount of information on the internet and reporting it back to its user. However, not everything online is accurate, and you may not always know the source of data or information you receive.

Causey says the best method for anyone using information from tools like ChatGPT is “sandwiching,” which he describes as a process of providing the prompt to the AI, receiving and editing the response, sending the edited information back for more suggested revisions and so on.

“It's not replacing the work that you're doing or the ideas you're having,” Causey says. “It's like you're getting another set of eyes on it, like sending it to a colleague.”


Image is of Trey Causey. He is a middle aged, white man with graying hair that is darker on top and lighter on the sides, and graying facial hair. He is wearing glasses and a maroon colored polo shirt. He is standing with his hand behind his back in a conference room. There is a wall sized notepad on one side and a white board on the wall behind him.
Trey Causey, head of AI ethics and a data science director at Indeed: “Be very skeptical of vendors who claim to have solved hard problems with AI, or who think that they can magically do your hiring process by AI."

But AI isn't just for employers. Job seekers can use it too, and recruiters need to be on the lookout for cover letters and resumes that a job seeker may have created with AI, which could be used to exaggerate or misrepresent the applicant’s experience. Recruiters should be wary of monotonous content coming from job seekers that could indicate the use of AI. Companies have created tools in attempts to detect AI-generated content, including a free one released by ChatGPT’s creator, but these are not particularly effective — OpenAI estimates their tool does not catch 74% of AI-generated content.

AI Can Help Create Job Titles, Descriptions and Messaging Frameworks

In a survey from our Indeed Insiders — a community of customers of small to large businesses who use Indeed regularly and provide product feedback — recruiters reported mainly using AI and ChatGPT in the early stages of recruitment. They described these tools as helpful in:

  • Improving language used on career sites, job boards and social media
  • Developing job titles and descriptions, as well as a description of the hiring company
  • Creating boolean keyword recommendations 
  • Mapping out recruitment strategy/plan templates
  • Conducting research on finding a rare skill set
  • Writing email templates for communicating with candidates
  • Creating messages to attract job seekers 

One Insider says, “It makes life really easy. I am not starting from scratch with writing, and I just need to spend time refining and tweaking ChatGPT’s recommendations.”

AI Can Help Reduce Bias in Screening — to an Extent

Harvard Business Review says human bias can influence AI as it has been created by humans and mimics our decision making. This could be a concern with anything created by AI, such as job descriptions. From their experience using AI in recruitment, an Indeed Insider says, “[AI] won't [understand] that putting in a qualification like ‘you must have graduated from Stanford in the last 15 years’ is age discrimination.”

While efforts are being made to improve this aspect of ChatGPT and modern AI, the tools aren’t currently intelligent enough to understand nuances like educational, economic and other barriers that limit job seekers’ opportunities in the workplace. A critical review from a recruiter — not only knowledgeable in the limitations of AI but also in skills-based hiring — would be useful to supplement automation tools.

According to Recruiter.com, AI can reduce bias in screening candidates by focusing only on qualifications and skills, while ignoring factors like gender and ethnicity. ChatGPT can be used to efficiently conduct research, as engineers at the University of Wisconsin–Madison have found, so it could also be used to check that you offer fair base salaries for your roles. Consistently reviewing your pay processes is important to ensure pay equity within your company and especially critical as pay transparency laws become more widespread.

AI Can Give You More Time to Spend With Candidates and New Hires

Those surveyed in the Indeed Insiders community say one of the greatest benefits AI can deliver is changing where their time is spent by performing the administrative tasks they’d normally have to do themselves. Insiders report that decreasing the administrative workload helps them get to the meat of their job: recruiting, interviewing and onboarding new hires.

Image is of a woman sitting at her laptop taking notes, while on a virtual meeting. The main focus of the image is the man on the laptop screen. He is a black man with short, black hair and facial hair. He is sitting in front of a bookshelf resting his elbows on the desk. He appears to be laughing at the camera and wearing glasses, and a polka dot print dress shirt. He has a black sport watch on his right hand.
One Indeed Insider says, “The most exciting aspect for us is figuring out how we can use AI to automate tasks and take care of things in the background so that we can spend more time with candidates.”

“The most exciting aspect for us is figuring out how we can use AI to automate tasks and take care of things in the background so that we can spend more time with candidates and actually recruit, … talking to people and being more involved,” says one Indeed Insider.

Like other HR technology, AI can be an effective tool for streamlining workloads and getting to the invaluable experience of human-to-human interactions.

AI Can’t Replace the “Human” in Human Resources 

While tools like ChatGPT can mimic conversation and streamline tasks like writing communication templates, it’s not the same as the human approach to recruiting and hiring. This includes assessing a candidate’s potential, the sensitive task of rejecting candidates and guiding onboarding processes for new hires.

“You need to have human interaction with candidates to determine whether they are going to be the right fit to deal with clients and have nuanced problem-solving skills,” says one Indeed Insider. “I can't imagine trusting a machine to make a hiring decision like this.”

While some Insiders expressed concerns about losing the personal touch and authenticity to their content, or content becoming too homogeneous, others say they value AI as a tool for sourcing ideas and data. “I don’t see it replacing anything our recruiters do,” says another Insider. “It’s just an additional way of doing things and an interesting way to compare ideas.”

However, Causey cautions against embracing AI like ChatGPT without proper human oversight.

“Be very skeptical of vendors who claim to have solved hard problems with AI or who think that they can magically do your hiring process with AI,” he says. In addition, be wary of sharing sensitive information, such as turnover, compensation and personal employee data, with unvetted AI platforms. 

In light of a recent data breach that allowed ChatGPT users to see others’ messages, numerous countries are examining how to regulate AI technology. In a petition signed by over 1,000 AI experts, AI developers are being asked to pause further work on AI past GPT-4 (the latest version of ChatGPT) until safety protocols are put into place. For companies deploying AI, Causey advocates for a governance model that brings together individuals from different interdisciplinary backgrounds — in addition to the programmers — who can provide impartial oversight and objectively consider the applications for end users in order to responsibly deploy the new technology. 

In his view, we may not have to fear the apocalyptic scenario of robots taking over, but we should approach these new technological developments with cautious optimism. “It's like any other tool,” he says. “It's just an efficiency gain.”