Indeed’s ghost resumes

Jim Durbin of Respondable Recruitment Marketing is known by those in the job board and other talent acquisition technology communities as the Indeed Whisperer.

He’s a frequent and welcome attendee of the monthly Job Board Leaders’ Roundtable calls hosted by Steven Rothberg of College Recruiter job search site. Today, Jim was our guest.

We discussed the impact AI is having on Indeed, other job boards, and the entire recruitment industry. The benefits are many and significant. But so are the problems. Some lump all into one bucket and call it “fake resumes”. And, some are from fake people committing corporate espionage. Others, however, are more akin to AI slop as they represent real people applying to real jobs, but do a terrible job of communicating the credentials of the candidate.

Are the solutions easy? No, but our group today discussed several that should help the industry make progress, even if the outcomes are not perfect.

AI-generated summary:

Leaders explored candidate data quality and identified how artificial intelligence negatively impacts recruitment and database integrity.

Defining Ghost Resume Challenges
Participants categorized deceptive candidate data into malicious actors, low-intent applicants, and AI-generated design failures. These issues collectively poison database integrity and hamper effective recruitment.

Impact of AI Standardization
Excessive reliance on AI resume tools creates generic content that hides talent and prevents recruiters from identifying unique candidates. This technological uniformity undermines human nuance in hiring.

Proposed UX Solutions
Adding manual friction to application workflows can improve candidate intentionality and data quality. Prioritizing verified, high-intent data over high-volume models is essential for future hiring success.

Details

• Meeting Introduction and Topic Overview: Steven Rothberg opened the July 2026 job board leaders roundtable by introducing the session’s focus on candidate data quality, specifically the prevalence of “fake” versus “real” candidates within the industry. Rothberg introduced Jim Durbin, referred to as the “Indeed Whisperer” due to their extensive knowledge of the Indeed platform, to lead the discussion on how to categorize and address the issue of ghost resumes (00:00:08) (00:02:17).
• Jim Durbin’s Professional Background: Durbin provided an overview of their 25-year career in staffing, recruiting, headhunting, and corporate RPO. Durbin noted that their deep technical understanding of platforms like Indeed was gained through hands-on experience managing millions of dollars in advertising spend for major clients, which eventually led them to consult on internal database integrity and job board operations (00:04:26).
• Defining the “Ghost Resume” Problem: Durbin argued that the industry lacks a standardized definition for “ghost resumes,” causing confusion among stakeholders. Durbin emphasized that simply labeling everything as fraud is inaccurate, noting that the issue encompasses multiple distinct categories of candidate behavior and platform-wide data quality problems that need to be addressed internally by job boards rather than dismissed (00:06:02).
• Category 1: Malicious and Adversarial Actors: Durbin identified the first category of fake candidates as malicious actors, specifically citing corporate espionage and sophisticated hacking attempts by groups like North Korean bad actors who use deep fakes, stolen imagery, and AI-assisted communication to gain access to corporate systems (00:07:37). These threats require high-level operational security, with Durbin noting that some organizations are even employing countermeasures like requiring candidates to physically perform tasks on camera to verify their identity (00:08:41).
• Category 2: Poorly Sourced, High-Volume Applicants: Durbin defined a second category as real candidates who apply to hundreds of jobs using automated tools without genuine intent or context. After auditing 200 applicants on Indeed, Durbin found that every candidate contacted was a real person, though they were often unqualified or applying indiscriminately, resulting in a significant portion of employer budgets being wasted on high-volume, low-intent applications (00:11:13).
• Category 3: The “XX” AI Resume Issue: Durbin highlighted a specific design flaw in the Indeed platform where the user interface suggested AI-generated bullet points that resulted in 750,000 resumes (as of January) containing placeholder text like “XX%” instead of actual data (00:13:43) (00:16:49). Durbin characterized this as a design failure rather than solely an AI problem, noting that once this data was injected into the database, it was difficult for the platform to revert (00:14:34) (00:35:46).
• Impact on Database Integrity: Durbin warned that the proliferation of AI-altered resumes has “poisoned” databases, creating a “mushy middle” where candidates look identical, making it increasingly difficult for recruiters to identify top talent (00:14:34) (00:17:42). Durbin estimated that approximately 50% of profiles now exhibit some form of AI adjustment, which makes it challenging for sorting algorithms and recruiters to distinguish between qualified and unqualified individuals (00:17:42) (00:35:07).
• The Rise of AI-Assisted Resumes and Standardization: Durbin and Rothberg discussed how candidates are using AI to rewrite their resumes to match job descriptions, leading to a loss of unique “voice” in applications (00:23:51). Durbin noted that because candidates are outsourcing their cognitive thinking to AI, they are becoming less creative and more prone to submitting standardized, generic content that recruiters find impossible to differentiate (00:24:34) (00:32:45).
• Interview Questions and “Sameness”: Paul Toomey noted that the issue of standardization extends beyond resumes to interview questions, as both interviewers and candidates are relying on the same AI tools to generate content. This creates a cyclical problem where everything sounds the same, and the lack of human nuance makes the recruitment process feel like a game rather than a professional assessment (00:27:44) (00:29:13).
• Cognitive Impact and Candidate Competence: Durbin observed that reliance on AI tools is causing candidates to feel “dumber and more arrogant” because they believe they have successfully “beaten the system” without actually doing the work (00:31:51). Durbin argued that this outsourcing of thinking harms candidates’ ability to perform in actual jobs, as they fail to learn the underlying skills required for the roles they are applying for (00:32:45) (00:48:55).
• Recruiter Differentiation Challenges: Rothberg and Durbin discussed the loss of “nuggets” of talent in the current system (00:40:29) (00:50:35). Durbin explained that in the past, recruiters could look at hundreds of resumes and identify patterns or help candidates rewrite poor resumes to showcase their actual skills; however, the current volume and AI-generated uniformity have stripped recruiters of the ability to spot these patterns, making the manual assessment of candidates much more time-consuming (00:51:45).
• UX Design as a Solution: Alex Murphy suggested that the solution lies in platform design, specifically by introducing friction—such as short codes or manual validation steps—into the application workflow to force candidates to be intentional. Murphy argued that job boards must design “not happy path” scenarios to prevent users from blindly hitting an “easy button” that leads to templated, low-quality content (00:53:18).
• Equity and Fairness in Hiring: Paul Toomey raised the point that AI is technically “equitable” because it provides resume-writing tools to those who previously could not afford them, but this has backfired by flooding the market with generic, high-quality-looking resumes (00:46:57). Durbin countered that while access is a factor, the real goal of recruiting is selection and talent identification, not just “fairness,” and that the current model is failing to identify the best candidates (00:48:06).
• Future Outlook and Market Dynamics: The participants concluded that candidates have been trained to “game” the system because they receive no feedback or response from the majority of employers (00:55:54). Durbin noted that until recruiters and job boards solve the issue of candidate feedback and prioritize high-intent, verified candidate data, the current “volume” model will continue to degrade the quality of hiring, forcing recruiters to rely on more intensive, high-touch methods to find legitimate talent (00:55:07).
• Industry Employment Challenges: Steven Rothberg discussed an industry-wide problem regarding the hiring process, characterizing it as a “tax on employment” that makes it increasingly difficult for employers to hire and for candidates to find suitable roles. They expressed a belief that the industry is still in a position to correct this course (00:57:51).
• Upcoming Meeting and Topic: Steven Rothberg announced that the next meeting will be held on the second Thursday of August, featuring guest speaker Yaz Dalal from Jovio. The session will address the concepts of “post and pray” versus Return on Investment (ROI) in hiring, with Steven Rothberg noting that these are distinct strategies (00:57:51).
• Contact Information for Jim Durbin: Jim Durbin advised that the best way to make contact is through LinkedIn or via the website responsible.io. Jim Durbin provided the LinkedIn profile identifier as “Jim Durbin” and confirmed the website address is responsible.io (00:57:51).