For employers, the second issue is becoming harder to ignore.
The Institute of Student Employers recently reported that two-thirds of employers believe graduates and apprentices are using AI to misrepresent their skills during selection, up from around half in 2025. The same reporting found that employers expect AI to reshape entry-level roles, with the focus shifting towards evolving tasks rather than simply removing jobs.
That creates a difficult screening problem. Graduate candidates often have limited work experience, so recruiters have traditionally relied on signals such as academic background, written answers, CV quality and interview performance. If those signals can be heavily polished or partially produced by AI, they become less useful as evidence of genuine capability.
Why traditional screening is under pressure
A well-written application used to (allegedly) tell recruiters something about effort, communication and motivation. It was never a perfect measure, but it had some value.
Now, a candidate can use AI to rewrite answers, tailor a CV to a job description or practise polished responses. Some of this is reasonable preparation. But when AI starts masking gaps in reasoning, communication or self-awareness, hiring teams risk making decisions on presentation rather than potential.
This is especially important in graduate hiring, where the real question is rarely “has this person done the job before?” It is usually “can this person learn, adapt and perform with support?”
What AI-resistant assessment means
AI-resistant assessment does not mean trying to detect every use of technology. That is unlikely to be realistic and damaging to the candidate experience.
So what makes a test ai-resistant? A better approach is to design selection methods that make external assistance less useful. This can include interactive assessments, time-bound tasks, adaptive question formats, structured scoring and exercises that measure reasoning or judgement rather than rehearsed written content.
Used well, these gamified assessments give recruiters a clearer view of how candidates think, not just how well they can package themselves.
A more balanced graduate process
Graduate recruitment still needs human judgement. Interviews, candidate conversations and contextual information all matter. But relying too heavily on CVs and written applications is becoming riskier than it already was.
The strongest processes will combine candidate-friendly assessment with structured human review. That gives employers better evidence, while still giving graduates a fair chance to show potential.
As AI becomes a normal part of applying for jobs, the question for employers is no longer whether candidates will use it. The question is whether the hiring process can still identify real ability when they do.
