Almost exactly one year ago, What They Think published an article titled "What AI Means for Your Future Workforce." It opened like this:

While tremendously useful for things like churning out certain types of marketing copy, analyzing customer data, and writing code, Gen AI is also deeply impacting the future workforce by churning out impatient, superficial thinkers. AI enables students to breeze through tasks that once required deep reading, analysis, and revision, all without learning a thing.

Today, all of the lessons of that article are showing up in hiring surveys. The Hirevue Early Careers Hiring Global Benchmark Report, released in 2026, for example, synthesizes data from the Institute of Student Employers (UK), the National Association of Colleges and Employers (US), and the Australian Association of Graduate Employers. The picture it paints is one of employers scrambling to adjust hiring practices because traditional proxies for competence—degrees, transcripts, resumes—no longer tell them anything about what candidates can actually do.

The Cheating Problem

In the UK, 61% of employers report that candidates are using AI during interviews without disclosure. That number isn’t surprising when you consider that 79% of Australian higher education and supplier sectors believe students are using GenAI significantly in their recruitment efforts. The report notes that suspected cheating in assessments has more than doubled, from 7% to 15% of employers who “frequently” encounter it.

The response from employers? Seventy-nine percent of UK employers are now “either redesigning or reviewing their recruitment processes because of AI developments.” A hiring “arms race” of sorts. In Australia, 15% of employers have adjusted their application or selection processes in response to candidate GenAI use, a jump from just 2% the previous year.

But employers aren’t just worried about cheating. They’re worried because they’ve lost faith in the traditional signals that were supposed to tell them who was competent.

The Death of Credentials

For years, a degree was the answer. In the UK, for example, a 2:1 (upperclass honors degree) from a decent university meant something. Today? UK employers have slashed the requirement for a 2:1 degree from 71% of organizations in 2014 to 47% today. In the US, while 79% of entry-level jobs still technically require a bachelor's degree, employers are de-emphasizing GPA as a screening tool. They’ve realized that when every student has access to ChatGPT, a transcript tells you nothing about what that person can actually think.

The report itself acknowledges this directly. According to Dr. Mike Hudy, Chief Science Officer at Hirevue: “Traditional proxies like résumés summarizing past experiences hold minimal predictive power, as candidates often lack a substantive professional history to review.”

What he’s really saying: We can’t trust anything you’ve told us about yourself because we have no way to verify it was actually you who did the work.

The Skills Gap That AI Created

So what are employers now desperately looking for? Communication, teamwork, and critical thinking. These—deep reading, analysis, complex problem-solving, and the ability to think through something without immediately outsourcing it to a machine—are the very skills that AI is eroding.

Employers have been forced to adopt skills-based hiring because credentials stopped meaning anything. McKinsey research cited in the report found that skills-based hiring is five times more predictive of job performance than hiring for education. That’s not because skills-based hiring suddenly became better. It’s because education became useless as a signal.

Seventy percent of US employers now report using skills-based hiring practices. They're not doing it because they love skills assessments. They're doing it because they have no other choice. A resume could be AI-generated. A transcript could represent work done by ChatGPT. The only thing left to assess is what someone can actually do in real time, under observation.