Pre-Employment Assessments in the Age of AI

‍

Some candidates are exceptionally good at getting hired. That does not necessarily mean they will be exceptionally good at the job. They know how to build a polished résumé. They communicate confidently. They tell compelling stories. They anticipate interview questions and understand what hiring managers want to hear.

‍

Those skills have always influenced hiring decisions. But now, AI can make almost anyone better seem better on paper or even help coach them through the interview. That creates a growing challenge for talent acquisition leaders and hiring managers: Are we measuring a candidate's potential to perform the job, or their ability to present themselves as someone who can? The distinction matters because interview performance is not the same thing as job performance.

‍

And as AI transforms résumé screening, candidate preparation, and recruiting, organizations may need to rethink what they focus on for hiring, what deserves the most weight, and where validated pre-employment assessments can provide stronger evidence of potential job performance.

‍

TLDR

‍

AI is making it easier for candidates to create polished resumes, applications, and interview responses, making presentation alone a less reliable indicator of future job performance. Employers can strengthen hiring decisions by combining structured interviews with validated pre-employment assessments, situational judgment tests, work samples, and other job-relevant measures. The goal is not to remove human judgment from hiring, but to give hiring managers stronger, more objective evidence of a candidate’s potential to perform.

‍

Hiring Has Always Had a Presentation Problem

‍

Long before ChatGPT and Claude were on the scene, hiring managers could confuse confidence with competence, charisma with capability, and likability with job fit. Selection research has documented this problem for decades. A particularly important meta-analysis by Barrick et al. (2009) examined impression management in employment interviews in comparison to future performance.  

‍

They found that candidates who were good at managing impressions had a substantial advantage in the interview, but that advantage translated far less to performance on the job. Impression management was associated with about 22% of the differences in interview ratings (r = .47,), compared with only about 2% of the differences in later job performance (r = .15).

‍

In plain English, candidates who were particularly effective at managing how interviewers perceived them tended to receive considerably better interview evaluations. That advantage carried over much less strongly to how they actually performed at work. This does not make interviews useless. It means employers need to distinguish between what interviews can reveal about candidates and what candidates can influence about interviews.

‍

Structured Interviews Produce Better Hiring Signals

‍

The problem becomes larger when every hiring manager asks different questions, follows different instincts, and decides what "good" looks like after meeting the candidate. In another important large meta-analytic study, McDaniel et al. (1994), covering more than 86,000 individuals, found that interview validity varied considerably based on how the interview was designed, with structured approaches producing stronger job-performance prediction than unstructured interviews.

‍

More recent research reinforces that conclusion. Sackett et al. (2022) re-evaluation of personnel selection research placed structured interviews among the strongest available predictors of job performance. Their estimated validity of r = .42 means structured interview scores shared about 18% of the variance with later job performance, a substantial relationship for any single hiring method.

‍

There is also a broader lesson here about gut instinct. Kuncel et al. (2013) examined what happens when employers have essentially the same candidate information, but combine it in different ways. One approach used holistic human judgment. The other applied predetermined rules to combine information such as interviews, assessments, and other candidate data.

‍

For predicting job performance, mechanical combination produced an average validity of .44 compared with .28 for holistic judgment, an improvement of more than 50% in predictive validity. This was not an argument for letting today's AI systems make hiring decisions. It showed something simpler: consistently applying the same evidence can outperform reviewing everything and relying on overall gut feel. Experience still matters. Human judgment still matters. But "I'll know the right candidate when I meet them" is not a selection strategy.

‍

AI Is Amplifying the Presentation Problem

‍

What has changed is the scale. According to LinkedIn's 2026 recruiting research, U.S. applications per open role had doubled since spring 2022, while 66% of recruiters said finding qualified talent had become harder. LinkedIn also reported that 93% of recruiters planned to increase their use of AI in 2026. (LinkedIn Corporate Communications, 2026)

‍

LinkedIn’s 2026 recruiting research

‍

Candidates are using AI too.

‍

Gartner reported that 39% of candidates surveyed had used AI during the application process. Among those AI users, 54% used it to generate résumé or CV content, 50% used it for cover letters, and 29% used it to answer questions on assessments. The last point is making the use of un-proctored skills and cognitive tests much more difficult to rely on. (Gartner, Inc., 2025)

‍

Gartner’s candidate AI-use research

‍

Some of this is completely legitimate. Helping someone communicate real experience more clearly is not the same as deception. But AI-assisted polishing and AI-assisted fabrication are different problems. Long before generative AI, Henle et al. (2019) identified three distinct forms of intentional résumé deception: fabrication, making information up; embellishment, exaggerating information that is true; and omission, deliberately leaving out relevant information. In a sample of 262 recent job seekers, 34% reported at least some fabrication, 66% some embellishment, and 62% some omission. These figures do not mean that most résumés are fraudulent, but they do show that intentional résumé distortion is not uncommon.

‍

AI did not create these behaviors. What AI changes is how quickly and easily someone can turn exaggerated or fabricated claims into polished, job-specific language. Consider a job description asking for someone who is customer-focused, detail-oriented, proactive, comfortable managing multiple priorities, and skilled at cross-functional collaboration.

‍

A candidate can now ask an AI system: "Rewrite my professional summary so I appear to strongly demonstrate these five qualities." Almost instantly, the résumé may sound remarkably similar to the employer's description of its ideal candidate. If another AI system then evaluates that résumé against the original job description, hiring risks becoming AI evaluating how effectively another AI reproduced the employer's language. Language alignment is not necessarily evidence of capability.

‍

Better Writing Really Can Change Hiring Outcomes

‍

This is not just hypothetical. A field study by Wingate et al. (2025) followed 183 Canadian co-op students through their job searches. Better written, clearer, and better structured résumés and cover letters were associated with more interview opportunities and faster placement, even after controlling for experience, accomplishments, and job tailoring. Because the study was observational, it does not establish that writing quality alone caused those outcomes. It does show that presentation quality carries meaningful weight.  

‍

Even stronger causal evidence comes from a large field experiment by Wiles et al. (2025) involving nearly half a million jobseekers. Applicants given non-generative algorithmic writing assistance produced résumés with fewer errors and were ultimately hired about 8% more often, at wages roughly 10% higher.  

‍

And new 2026 Management Science research by Cowgill et al. (2026) found that when people being evaluated had access to generative AI, evaluators' screening accuracy fell by roughly 4% to 9% in some settings. AI helped less-qualified people produce communication that more closely resembled that of experts, although effects varied across contexts.  

‍

That is the deeper AI in hiring problem. The relationship between how qualified someone sounds and how qualified they actually are may be getting weaker.

‍

Move Hiring From Presentation Signals to Performance Signals

‍

The solution is not eliminating résumés or interviews. It is changing the mix of evidence. Think of candidate information as a radio signal. Résumé polish, charisma, keyword optimization, confidence, and storytelling can create a lot of volume. What employers need is a stronger signal underneath the noise.

‍

Presentation signals include:

• Résumé wording  

• Professional summaries  

• Interview polish  

• Confidence and charisma  

• Storytelling ability  

• Keyword optimization  

‍

Performance-oriented signals can include:

• Job-relevant behavioral tendencies  

• Situational judgment  

• Work samples  

• Structured interview evidence  

• Problem-solving demonstrations  

• Validated pre-employment assessment results

‍

One question every hiring team should ask is: What evidence in our hiring process exists independently of the candidate's ability to describe themselves? That is signal verification.

‍

Can Candidates Fake Personality Assessments For Hiring?

‍

Yes, candidates can attempt to manage impressions on personality assessments. That limitation should be acknowledged rather than dismissed. But the more important scientific question is whether personality assessments continue to meaningfully differentiate candidates and predict relevant outcomes, despite this.

‍

Hogan et al. (2007) examined 5,266 real applicants who had previously been rejected and later reapplied for the same job. These candidates had a clear incentive to improve their results. On any individual personality scale, 5.2% or fewer showed the type of substantial increase identified by the researchers, and only three applicants (0.0006%) changed substantially across all five major scales beyond the study's statistical threshold.  

‍

That study does not prove personality assessments cannot be manipulated. No single study could. But it is useful real-world evidence that completely transforming a multidimensional personality profile appears considerably harder than simply knowing what an employer wants and inserting those qualities into a résumé.

‍

Complementary evidence comes from a meta-analysis by Ones et al. (1996). They found that socially desirable responding did not explain away the relationship between personality measures and job performance. In other words, even when candidates may be trying to present themselves favorably, personality assessments can still provide meaningful information related to future job performance. The practical question should therefore be: Does this assessment continue to predict outcomes that matter? That is the standard validated hiring tools should be held to.

‍

Situational Judgment Tests Shift the Question From “Who Are You?” to “What Would You Do?”

‍

Situational judgment tests, or SJTs, take another approach. Rather than asking candidates to say, "I have great customer judgment," an SJT gives them a realistic customer situation and asks what they would do. Instead of asking candidates to describe good judgment, the assessment gives them an opportunity to demonstrate judgment.

‍

There is substantial evidence that this matters. A meta-analysis by McDaniel et al. (2001) covering more than 10,000 participants found that SJT scores meaningfully predicted job performance. Later research found that SJTs can also add incremental prediction of job performance beyond cognitive ability and personality assessments (McDaniel et al., 2007). In one particularly compelling longitudinal study, Lievens and Sackett (2012) followed 723 individuals from admission into their careers and found that an interpersonal SJT predicted internship performance seven years later and job performance nine years later.

‍

SJTs are not immune to AI or coaching. But emerging research suggests that well-designed SJTs may make AI assistance less valuable. Harwood et al. (2024) examined more than 107,000 real applicants taking a timed and virtually proctored SJT and found little change in performance after ChatGPT became widely available. In a second experiment, participants who were explicitly instructed to use ChatGPT achieved only slightly higher scores than those using other resources or no outside assistance. For employers, this strengthens the case for job-specific SJTs: rather than asking candidates to tell you they possess good judgment, give them realistic situations and collect evidence of how they apply it.  

‍

A Better Hiring Process Uses Multiple Independent Signals

‍

The strongest approach is not résumé versus assessment versus interview. It is a sequence in which each method adds different evidence: Résumé and background → Personality/SJT assessment → Structured interview → Integrated hiring decision

‍

The résumé establishes qualifications and career history. A job-relevant personality assessment and SJT add standardized information about behavioral tendencies and judgment. The structured interview then lets managers investigate experience, request concrete examples, explore discrepancies, and probe areas identified by the assessment. This is where hiring assessments can make interviews better, not replace them. A good assessment can tell the interviewer where to dig. A structured interview can then test those hypotheses with behavioral evidence.

‍

No personality assessment, SJT, structured interview, work sample, résumé, or AI model should make the entire hiring decision alone. Poor assessments do not become good simply because they are standardized. Selection tools need to be job relevant, appropriately validated, and combined thoughtfully. The goal is not removing human judgment from hiring.

‍

Frequently Asked Questions

‍

Can pre-employment assessments actually predict how someone will perform on the job?

‍

Yes, when they are designed and validated appropriately. Pre-employment assessments can measure job-relevant characteristics such as behavioral tendencies, cognitive abilities, situational judgment, and skills. The most effective assessments are tied to the requirements of the specific role and validated against real performance outcomes.

‍

Can candidates fake their answers on personality assessments?

‍

Candidates can try to present themselves in a more favorable way, but that does not necessarily eliminate an assessment’s ability to predict job performance. Research suggests that completely changing a multidimensional personality profile is more difficult than simply tailoring a resume or interview response to what an employer wants to hear. Personality assessments should also be used alongside other hiring methods rather than as the sole basis for a decision.

‍

Are structured interviews really better than traditional interviews?

‍

Research generally shows that structured interviews are more consistent predictors of job performance than unstructured interviews. Asking candidates the same job-relevant questions and evaluating their responses using predetermined criteria makes it easier to compare candidates consistently and reduces reliance on gut instinct.

‍

What does a situational judgment test actually tell you about a candidate?

‍

A situational judgment test gives candidates realistic workplace scenarios and asks how they would respond. Instead of asking someone to simply say they have good judgment or problem-solving skills, it provides insight into how they approach situations and decisions they may actually encounter on the job.

‍

Which pre-employment assessments are most useful when making a hiring decision?

‍

There is no single assessment that is best for every job. The right approach depends on the role and the characteristics that actually drive performance. Depending on the position, employers may use personality assessments, situational judgment tests, cognitive measures, skills assessments, work samples, or a combination of methods.

‍

How can you tell who the best candidate really is when everyone is using AI?

‍

The key is looking beyond how well a candidate presents themselves. Resumes, cover letters, and interview responses can increasingly be polished with AI, so employers need additional evidence of a candidate’s potential to perform. Combining job-relevant assessments, situational judgment tests, structured interviews, work samples, and experience gives hiring teams multiple independent signals to make a more informed decision.

‍

It is giving hiring managers better evidence to judge.

‍

If your hiring process still relies heavily on résumés, candidate presentation, and unstructured interviews, consider where stronger independent signals could improve the decision. ForPsyte's pre-hire assessments combine job-relevant assessment data with structured interviewing and validation to help organizations focus hiring decisions more directly on evidence connected to performance.

‍

For another perspective on identifying the signals that actually predict success, read Moneyball for Hiring.

‍

References

‍

Barrick, M. R., Shaffer, J. A., & DeGrassi, S. W. (2009). What you see may not be what you get: Relationships among self-presentation tactics and ratings of interview and job performance. Journal of Applied Psychology, 94(6), 1394–1411. https://doi.org/10.1037/a0016532

‍

Cowgill, B., Hernández-Lagos, P., & Wright, N. L. (2026). Does AI cheapen talk? Theory and evidence from global entrepreneurship and hiring. Management Science. Advance online publication. https://doi.org/10.1287/mnsc.2024.07027

‍

Gartner, Inc. (2025, July 31). Gartner survey shows just 26% of job applicants trust AI will fairly evaluate them. https://www.gartner.com/en/newsroom/press-releases/2025-07-31-gartner-survey-shows-just-26-percent-of-job-applicants-trust-ai-will-fairly-evaluate-them

‍

Harwood, H., Roulin, N., & Iqbal, M. Z. (2024). “Anything you can do, I can do”: Examining the use of ChatGPT in situational judgement tests for professional program admission. Journal of Vocational Behavior, 154, 104013. https://doi.org/10.1016/j.jvb.2024.104013

‍

Henle, C. A., Dineen, B. R., & Duffy, M. K. (2019). Assessing intentional resume deception: Development and nomological network of a resume fraud measure. Journal of Business and Psychology, 34(1), 87–106. https://doi.org/10.1007/s10869-017-9527-4

‍

Hogan, J., Barrett, P., & Hogan, R. (2007). Personality measurement, faking, and employment selection. Journal of Applied Psychology, 92(5), 1270–1285. https://doi.org/10.1037/0021-9010.92.5.1270

‍

Kuncel, N. R., Klieger, D. M., Connelly, B. S., & Ones, D. S. (2013). Mechanical versus clinical data combination in selection and admissions decisions: A meta-analysis. Journal of Applied Psychology, 98(6), 1060–1072. https://doi.org/10.1037/a0034156

‍

Lievens, F., & Sackett, P. R. (2012). The validity of interpersonal skills assessment via situational judgment tests for predicting academic success and job performance. Journal of Applied Psychology, 97(2), 460–468. https://doi.org/10.1037/a0025741

‍

LinkedIn Corporate Communications. (2026, January 7). LinkedIn research: Nearly 80% of people feel unprepared to find a job in 2026, as two-thirds of recruiters say it’s harder to find quality talent. https://news.linkedin.com/en-us/2026/LinkedIn-Research-Talent-2026

‍

McDaniel, M. A., Hartman, N. S., Whetzel, D. L., & Grubb, W. L., III. (2007). Situational judgment tests, response instructions, and validity: A meta-analysis. Personnel Psychology, 60(1), 63–91. https://doi.org/10.1111/j.1744-6570.2007.00065.x

‍

McDaniel, M. A., Morgeson, F. P., Finnegan, E. B., Campion, M. A., & Braverman, E. P. (2001). Use of situational judgment tests to predict job performance: A clarification of the literature. Journal of Applied Psychology, 86(4), 730–740. https://doi.org/10.1037/0021-9010.86.4.730

‍

McDaniel, M. A., Whetzel, D. L., Schmidt, F. L., & Maurer, S. D. (1994). The validity of employment interviews: A comprehensive review and meta-analysis. Journal of Applied Psychology, 79(4), 599–616. https://doi.org/10.1037/0021-9010.79.4.599

‍

Ones, D. S., Viswesvaran, C., & Reiss, A. D. (1996). Role of social desirability in personality testing for personnel selection: The red herring. Journal of Applied Psychology, 81(6), 660–679. https://doi.org/10.1037/0021-9010.81.6.660

‍

Sackett, P. R., Zhang, C., Berry, C. M., & Lievens, F. (2022). Revisiting meta-analytic estimates of validity in personnel selection: Addressing systematic overcorrection for restriction of range. Journal of Applied Psychology, 107(11), 2040–2068. https://doi.org/10.1037/apl0000994

‍

Wiles, E., Munyikwa, Z., & Horton, J. (2025). Algorithmic writing assistance on jobseekers’ resumes increases hires. Management Science, 71(12), 10144–10164. https://doi.org/10.1287/mnsc.2024.04528

‍

Wingate, T. G., Robie, C., Powell, D. M., & Bourdage, J. S. (2025). The signals that matter: Resumes, cover letters, and success on the job search. International Journal of Selection and Assessment, 33(3), e70022. https://doi.org/10.1111/ijsa.70022