The Impact of AI on Talent Recruitment with Julia Medvin and Chelsea Stearns

HOST: Here we are, the 22nd episode of Talking Talent, joined by two HR leaders I have known for a long time, Julia Medvin and Chelsea Sterns. We will discuss the state of talent acquisition as it relates to artificial intelligence, how AI has shaped applying for jobs and resumes, and the pros and cons of candidates completing different types of assessments during the interview process.

HOST: We cover a lot of ground today. If you enjoy the episode, please share it, and you can subscribe anywhere you get podcasts, including Spotify and Apple, or on our website.

HOST: Today I am joined by two people I have known for a number of years and deeply respect for their work: Julia Medvin and Chelsea Sterns. Julia and Chelsea, thank you for joining me today.

JULIA MEDVIN: Thanks for having me. Excited to be here.

CHELSEA STERNS: Thanks for having me. Excited to be here.

HOST: Julia and Chelsea both have the same job title, but their career paths are different. Julia is also the first person to be on this podcast twice. Would you both give a quick highlight of your backgrounds? Julia, could you share first?

JULIA MEDVIN: Thank you for inviting me back. I am currently the vice president of human resources at NRU CFC, which is essentially a nonprofit financial institution focused on providing financial support and lending to rural electric cooperatives.

JULIA MEDVIN: My career has moved between unique organizations. I started in HR at Accenture after getting an MBA in finance and international business. The past five years I spent in the startup world helping companies stand up people operations. Prior to that, I spent about 10 years in consulting working with Fortune companies on strategies to attract, retain, and engage employees.

JULIA MEDVIN: I remain a numbers person, but HR is important to me because I appreciate the psychology behind employee engagement and loyalty, and how employers can create work environments that work for their people.

HOST: Chelsea, your background?

CHELSEA STERNS: My path is different but shares a focus on numbers. I earned an undergraduate degree in business management with a focus in finance, then started working in recruiting and stayed in recruiting for nearly a decade.

CHELSEA STERNS: I later returned to school for a master’s in human resources management from Georgetown, which helped me pivot into an HR generalist capacity. I am now vice president of human resources at Perry Labs, a government contractor with offices across the country.

CHELSEA STERNS: At Perry Labs we focus on digital systems integration, and much of my talent acquisition work has been in engineering recruitment. I also teach for SHRM, facilitating their talent acquisition specialty credential seminar and their internal workplace investigations specialty credential seminar.

CHELSEA STERNS: I teach a talent acquisition and management class for Georgetown’s master’s in human resources management program as well.

HOST: There are many paths to reach the roles you both hold. I also want to note that the notion HR people can’t do numbers is outdated. You both emphasized the importance of understanding quantitative data to be strategic HR leaders.

HOST: Today we will focus on the company side of talent acquisition and what you both are seeing in your organizations, especially related to AI, resumes, and candidate assessments. We’ll start with AI. From a macro level, what are you seeing and what are some pros and cons of AI in recruiting from both the company and candidate perspectives? Julia, I’ll start with you.

JULIA MEDVIN: It’s a learning experience as we see more about what AI can do and how people leverage it. One clear change is expectations around resume quality: writing and formatting have improved, and errors or misspellings are less tolerated now that these tools exist.

JULIA MEDVIN: The downside is personal character can be harder to see. Some resumes feel embellished—people with two years of experience claiming extensive technical skills or leadership and dramatic business impact. That makes it harder to determine what candidates have actually done.

JULIA MEDVIN: As a result, I need to test and vet candidates more carefully and ask better questions to differentiate who will be successful culturally and technically. Resumes used to help weed candidates out; now they often look so polished it’s harder to separate candidates who truly have the required skills and growth potential.

HOST: Remote video interviews have become commonplace for early interview steps, which affects how candidates present themselves. Chelsea, you’ve had experiences with ChatGPT and other AI. Can you follow up on what Julia said and share your insights?

CHELSEA STERNS: We are seeing resumes that are much better written, which is good, but it’s also harder to distinguish genuine experience from embellished claims. A dead giveaway that a document has been run through ChatGPT, in my experience, is the frequent use of an em dash. Historically we saw hyphens, but ChatGPT often inserts em dashes.

CHELSEA STERNS: We also see highly customized resumes created in seconds by providing a resume and job description to ChatGPT and asking it to tailor the resume for that role. Candidates can revamp their resumes instantly.

CHELSEA STERNS: I have had students tell me they will interview remotely with video on one half of the screen and ChatGPT on the other. When asked a question they do not know, they quickly use ChatGPT to produce an answer and regurgitate it to the interviewer. That is a real risk.

CHELSEA STERNS: Given these developments, the question becomes how we vet candidates more deeply to ensure we are getting the skills and experience we expect.

HOST: That is concerning, especially for small to mid-size companies that lack resources. If a resume looks polished but falls apart in front of a hiring manager, it becomes a company-wide responsibility to address vetting.

HOST: Julia, you mentioned masking or overselling skills on resumes. Chelsea, you described tailored resumes and candidates using ChatGPT during interviews. Is there anything else in resumes that stands out, or any advice for candidates structuring their materials?

JULIA MEDVIN: Resumes are increasingly embellished, and candidates claim extensive experience in short time frames. Talent acquisition teams, with hiring managers, must vet for required skills by using behavioral interview questions and digging into specific examples.

JULIA MEDVIN: For instance, if a candidate claims extensive experience in internal workplace investigations, ask for specific situations they investigated, how they documented it, and how they resolved it. Those examples validate the skills claimed on the resume.

CHELSEA STERNS: I agree. Resumes are much more customizable now, and embellishment is common. Behavioral interview questions that ask for examples of past application of skills help combat AI-generated embellishment.

HOST: It feels like in-person interviews are becoming more important because resumes are so polished. Even if you do vet online, a polished resume raises skepticism about authenticity. Julia, do you see in-person interviews regaining importance?

JULIA MEDVIN: Yes. In-person interviews are becoming more important for validation. Post-COVID we hired many people virtually and volume was manageable, but now resumes are too polished and sometimes embellished. In-person interviews, behavioral assessments, and validation of technical skills are becoming more necessary.

JULIA MEDVIN: For leadership roles, behavioral assessments are still relevant. For technical roles, assessments validating technical competency are increasingly important, even for junior candidates where basic skills like Excel proficiency may be overstated on resumes.

HOST: Chelsea, for distributed technical roles, such as software developers hired anywhere in the country, is there a way around bringing candidates to an office to test skills?

CHELSEA STERNS: You can make assessments timed and monitored, but that costs money. You can’t just send a coding prompt and expect it to be completed without using ChatGPT. To avoid AI assistance, you need proctored, timed assessments, which require investment and monitoring.

CHELSEA STERNS: Those solutions exist but increase cost and complexity. We are in a new world and learning how to navigate it.

HOST: There is also a candidate experience consideration—how many hoops do you require before they opt out? If you create too many barriers, you may lose candidates to quicker processes. The solution may be bespoke paths depending on role type: rapid hiring for high-demand technical skills with monitored assessments versus in-person and behavioral evaluations for leadership roles.

HOST: Let’s discuss assessments. I have seen predictive index assessments and in-person Excel and analytical assessments for HR roles. Julia, have you seen an increase in assessments, and are they specific to certain roles?

JULIA MEDVIN: Assessments have been around for a long time, but I am seeing more now. Historically, assessments were more behavioral for leadership roles and some technical tests for specific skills. Now, because resumes are less trustworthy and basic skills can be overstated, there is a greater need for technical assessments—even for junior roles—to validate skills like Excel and communication.

JULIA MEDVIN: Leadership assessments still matter for culture fit and leadership capability, but broader technical and foundational skill assessments are becoming more common.

CHELSEA STERNS: We have seen an uptick in simulations, case interviews, and cognitive and aptitude tests. Those assessments are useful and typically do a good job of evaluating the targeted skills. I find personality assessments to be less useful because they can be easy to game and may not provide actionable insights.

CHELSEA STERNS: Simulations and case interviews are valuable but expensive and time-consuming to administer and review, which increases time to fill. In defense and space engineering hiring, we keep time to fill under 25 days as a competitive advantage, so extensive assessments are used sparingly and mainly for leadership roles.

HOST: Assessments can also guide interviewers who are not HR experts. For senior roles, assessments can inform later interviewers about a candidate’s strengths. It’s important to communicate to candidates whether an assessment is pass/fail or primarily for guiding the interview process.

JULIA MEDVIN: I have mixed feelings about assessments, especially behavioral and aptitude tests. They can introduce subconscious bias if interviewers see results early. We prefer to save assessments for the end of the process and use them to inform the final decision.

HOST: Do applicant tracking systems have the ability to flag AI-generated signals, such as em dashes, before resumes reach recruiters? Is that a capability in current ATS systems?

CHELSEA STERNS: Some ATS platforms allow keyword searching and parsing, but not all systems flag AI-specific signals. Our HRIS/ATS is somewhat archaic and lacks that capability, but I assume more modern systems may have better detection functionality.

JULIA MEDVIN: Our current system is also older; we are moving to a newer system. Right now our focus is on finding the right skills quickly, not on automatically detecting AI use, although that question is valid and could be a product opportunity.

HOST: A Washington Post article cited Gartner research saying that by 2028 one in four job candidates globally could be fake due to AI-generated profiles. That is concerning.

HOST: Candidates should use AI as a tool in their toolkit to create resumes and present themselves, but not as a crutch. You must still be able to speak to your skills and experience.

HOST: Any final thoughts?

CHELSEA STERNS: No final thoughts beyond what we discussed. Thank you for having me.

JULIA MEDVIN: Thank you. I enjoyed the conversation.

HOST: When this episode is released, you can find Julia and Chelsea on LinkedIn. If you have questions for them, feel free to reach out to me as well. Thank you both for joining me.