HR professionals are confronting a scenario that is becoming increasingly familiar, and increasingly costly: A candidate’s resume looks perfect, the interview goes brilliantly, but once they are on the job, it becomes clear they lack the fundamental capabilities the role demands.
Alexander Alonso, Ph.D., SHRM-SCP
Chief Knowledge Officer, SHRM
At SHRM, we coined the term “skillfishing” to describe this phenomenon: the act of presenting credentials or capabilities that do not translate into real-world execution. While exaggerating skills is hardly a new problem, artificial intelligence has transformed a manageable challenge into a systemic crisis demanding a serious, structural response.
Polishing the bait
The modern job seeker increasingly turns to AI to navigate what many describe as a broken system, one in which automated screening tools filter out candidates before a human ever reviews their applications. To compete, candidates are leveraging AI in increasingly sophisticated ways.
- Automating the search: Bots now apply for high volumes of jobs on a candidate’s behalf, flooding recruiting pipelines with applications that may bear little relationship to real interest or fit.
- Building and tailoring resumes: Generative AI tools such as ChatGPT and Claude help candidates refine their professional history and over-optimize resumes with keywords mirroring job listings almost exactly, sometimes accurately and sometimes not.
- Crossing the line into embellishment: What begins as highlighting real strengths can shift into outright fabrication. In some cases, candidates use AI to generate work samples such as portfolios, writing, or code that they never actually produced, presenting them as evidence of capability they do not possess.
Why traditional interviews fail
Skillfishing is particularly dangerous because it exploits well-documented weaknesses in conventional hiring practices. Many employers continue to rely on a pair of 30-minute interviews prioritizing personality, cultural fit, and only shallow verification of claimed skills — precisely the conditions under which a skilled skillfisher thrives.
AI has enabled what some researchers are calling performance theater, where candidates convincingly demonstrate proficiency they do not actually have. In virtual environments, some applicants use invisible overlay tools to display AI-generated answers on their screens during live interviews, creating the appearance of working through a complex problem when they are, in effect, reading from a script. Recent data suggests that 38.5% of candidates show signs of cheating during interviews, a figure that rises to 48% for technical roles. These are not marginal numbers; they represent a structural challenge for any organization that has not yet adapted its assessment process to this new recruiting environment.
The cost of speed over signal
The organizational damage caused by a skillfisher rarely surfaces immediately. It typically takes 2-4 weeks for a manager to recognize a new hire who cannot independently tackle problems. By the time a formal performance conversation occurs, often 60-90 days into the role, the costs have already compounded across multiple dimensions:
- Financial loss: Accounting for recruitment, onboarding, lost productivity, and team disruption, replacing a bad hire carries a cost estimated at 50%-200% of the employee’s annual salary.
- Operational drag: Teams restructure around an underperforming hire, project timelines shift, and colleagues absorb additional workload, all consequences of a hire whose claimed skills were never verified against real-world demands.
Rebuilding the recruiting process
Addressing skillfishing requires more than incremental adjustments. Experts increasingly argue that organizations must undertake a fundamental redesign of how they evaluate candidates. The following approaches offer a practical framework for doing so.
1. Shift the metrics
Organizations that measure recruiting success primarily by time-to-fill are optimizing for speed at the expense of quality. A more meaningful set of indicators includes time-to-proficiency, or how quickly a new hire reaches independent competency, alongside quality-of-hire metrics such as 12-month retention rates and manager performance ratings. Reorienting around these signals encourages recruiters and hiring managers to make decisions that serve long-term organizational health, not just short-term pipeline velocity.
2. Implement multi-layered, role-specific assessments
Generic, off-the-shelf tests are among the easiest targets for AI-assisted candidates to game. A more defensible approach builds assessments that closely emulate the actual conditions of the role, drawing on situational judgment tests and ipsative, or forced-choice, instruments requiring contextual reasoning rather than the retrieval of memorized answers. Because these tools present candidates with competing priorities and real-world ambiguity, they are substantially harder for AI to navigate on a candidate’s behalf, and they reveal patterns of judgment a polished resume cannot.
3. Use AI as a defense
The same technology enabling skillfishing can also serve as a detection mechanism. AI-based interviewing and assessment platforms can identify anomalies human interviewers might overlook, including shortened response times, inconsistencies between assessment scores and responses to follow-up questions, or behavioral patterns suggesting scripted rather than spontaneous engagement. Deploying AI as a screening tool is not a replacement for human judgment; however, it adds a layer of verification candidates cannot easily anticipate or circumvent.
4. Incorporate real-world demonstrations
No assessment instrument substitutes fully for direct observation of a candidate’s work. Organizations prepared to invest in more rigorous evaluation have two particularly effective options. Trial projects allow hiring teams to observe a candidate’s ability to deliver outcomes under realistic conditions, treating early-stage hires, in effect, as freelancers who must demonstrate value before full commitment is made. In-person interviews, meanwhile, eliminate the technological advantages that virtual environments afford skillfishers. It is considerably more difficult to display competence you do not have when there is no screen to hide behind.
5. Assess for learning agility
As AI automates an expanding share of routine tasks, the most valuable capability a candidate can demonstrate is not necessarily what they know today, but how quickly and effectively they can develop new proficiency. Hiring processes that probe a candidate’s learning orientation, such as their track record of mastering unfamiliar tools, navigating ambiguity, and pivoting when conditions change, are better positioned to identify individuals who will remain effective contributors as the technological environment continues to evolve.
The proliferation of skillfishing may be a byproduct of the rise of AI, but I would counter that the remedy lies in employers’ strategic application of AI and in embracing the human elements of talent acquisition to discover real candidates with valued skills. The synergy of artificial intelligence and human intelligence will ultimately yield the highest return on investment of time, talent, and financial resources.
