When Everyone Can Look Qualified, What Becomes a Meaningful Signal?

A Signal Is Not the Same Thing as Capability

Generative AI is making it easier for candidates to produce polished résumés, tailored application materials, and highly prepared interview responses. At the same time, employers are reconsidering some of the proxies they have historically used to narrow large applicant pools, including GPA and degree requirements.

That does not mean traditional hiring signals have suddenly become useless. It does mean employers have a harder question to answer: what does a candidate signal actually tell us about the capability we are trying to assess?

As application polish becomes easier to produce and some traditional screening signals become less informative, employers need hiring evidence that sits closer to the capability itself—not to eliminate résumés, credentials, experience, or interviews, but to understand their limits.

Hiring has always relied on signals: a degree can signal education and persistence, a GPA can signal academic performance, prior employers can signal experience, an internship can signal exposure to a particular environment, and a polished résumé can signal communication and preparation.

But none of those things is the capability itself.

That distinction matters more as generative AI reduces the effort required to improve how a candidate presents those signals.

In its 2026 Job Outlook Spring Update, the National Association of Colleges and Employers found that 43.4% of surveyed organizations had detected AI-generated applications, while only 20.1% reported using tools to detect AI assistance during applications, assessments, or interviews.¹

Those figures do not tell us how many applications actually contain AI assistance or whether AI-assisted candidates perform worse after hire, but they do show that AI has become a visible part of the hiring environment before many employers have developed a consistent response.

More dramatic evidence comes from a 2026 Harvard Business Review article based on an operational study of 6,380 first-round remote screening sessions at U.S. technology companies. The authors reported suspicious behavioral patterns consistent with real-time assistance in nearly 60% of new-graduate software-engineering sessions and close to 40% of mid-level technical sessions, while rates remained below 10% in less technical roles.²

The authors are careful to note that these patterns were indicators requiring further review, not proof that individual candidates were cheating. They also explicitly disclose that they are technologists developing a startup around interview screening, making the findings better treated as an emerging operational signal than as evidence that the entire hiring market is experiencing the same phenomenon.

The useful conclusion is narrower: some hiring environments are becoming easier to optimize for without necessarily making underlying capability easier to see.

Employers Are Already Questioning Some Traditional Proxies

NACE has reported a significant decline in the use of GPA as an automatic screening mechanism in early-career recruiting, and its 2026 employer research found that the vast majority of employers prefer candidates to demonstrate skills through examples rather than simply list them.

Internship experience also continues to carry substantial weight in early-career decisions. In NACE’s Job Outlook 2026 research, having completed an internship with the hiring organization or within its industry were the only two candidate attributes employers rated as having “very much influence” when deciding between otherwise equally qualified candidates.³

That finding tells us what employers value. It does not establish that internship experience is universally the most predictive measure of future performance.

Degree requirements are changing too, although much more slowly than the rhetoric around skills-based hiring sometimes suggests.

A 2024 analysis from Harvard Business School and the Burning Glass Institute found that the annual number of roles from which employers removed degree requirements increased nearly fourfold between 2014 and 2023. Yet after accounting for occupational changes, the share of job postings requiring a degree declined by only 3.6 percentage points over the decade.⁵

Changing the wording of a job description is considerably easier than changing the decisions made later in the hiring process.

Removing a proxy does not automatically create a better signal.

Better Evidence Sits Closer to the Work

A 2023 reassessment of personnel-selection research published in Industrial and Organizational Psychology revisited the predictive validity of commonly used hiring methods. The updated estimates placed structured interviews at an operational validity of .42 and job-knowledge tests at .40, compared with .31 for general cognitive ability and .07 for years of job experience when predicting overall job performance.⁴

The authors describe this as support for the “sample” end of the sign-sample distinction: rather than primarily inferring capability from general signs, employers can design selection methods that sample behaviors connected to the work itself.

A résumé is largely a sign. Prior experience is a sign. A credential is a sign.

A structured discussion of how someone reasons through a job-relevant problem moves closer to a sample. So can an appropriately designed work exercise, job-knowledge assessment, or other opportunity to observe behavior connected to the role.

This does not make every work sample or structured interview automatically good—the same research shows substantial variation in validity across settings, and design, job analysis, standardization, and what candidates are actually asked to demonstrate all matter.

The implication is not simply “use more assessments.”

It is to ask a more disciplined question:

What evidence would actually help us observe the capability required for this role?

Skills-Based Hiring Fails When the Process Does Not Change

The Harvard Business School and Burning Glass Institute analysis examined more than 11,000 roles at large employers that had removed degree requirements. On average, those organizations increased the share of non-degreed hires in affected roles by only 3.5 percentage points. Because the changes applied to a relatively small share of roles overall, the researchers estimated that the movement translated into additional opportunities for roughly 97,000 workers out of approximately 77 million U.S. hires—less than one in 700 hires.

The firm-level patterns are even more revealing. The researchers classified 37% of companies as skills-based hiring leaders that produced sustained change, while 45% showed little meaningful change after removing degree requirements and another 18% initially changed before reverting toward previous hiring patterns.

Hiring managers still need a practical way to decide among candidates. When a familiar signal such as a degree disappears but nothing credible replaces it, decision-makers can fall back on the same proxy informally. The researchers argue that degrees can continue to function as risk-management heuristics when managers are under pressure to fill vacancies quickly.

Skills-based hiring therefore cannot stop at changing qualifications in the job description—the assessment process has to change with it.

Early-Career Hiring Requires a Different Standard

Employers also cannot always ask candidates to demonstrate work they have never had the opportunity to perform.

The 2023 reassessment of selection methods specifically notes that work samples and job-knowledge tests are often inappropriate for true entry-level hiring when the relevant knowledge and skills are expected to be developed after hire.

A new graduate should not necessarily be expected to demonstrate the same job-specific capability as someone who has already performed the role for five years. In those situations, relevant evidence may come from structured examples of problem solving, learning ability, judgment, communication, transferable experience, and the ability to acquire new knowledge.

General cognitive ability also remains relevant in this context. Although its estimated relationship with overall job performance has been revised downward in recent research, the same selection research notes that cognitive ability remains particularly useful in learning settings and roles where substantial post-hire training is central to performance.

The strongest hiring signal is therefore not necessarily the hardest assessment or the closest replica of the job—it is evidence appropriate to both the capability and the career stage being evaluated.

What Makes a Signal Meaningful?

The evidence suggests six useful questions employers can ask when designing or reviewing a hiring signal.

A meaningful signal should be:

Those questions become increasingly important when presentation itself becomes easier to manufacture.

The future of hiring is unlikely to be résumé-free, credential-free, or assessment-only. Employers will still need summaries, histories, qualifications, and efficient ways to narrow large applicant pools.

When almost everyone can look qualified, the competitive advantage in hiring will not come from finding a new proxy to replace the old one. It will come from designing a process that makes meaningful evidence of capability easier to see.