How To Hire Humans amid a Flood of AI Generated Resumes

AI is changing how candidates apply but small businesses can still hire well by testing for skills as part of the interview process, instead of poise and polish.

In just a few short years, generative AI has changed every aspect of how we do our jobs, including how we do them and how we, as employees, find them. For small businesses AI has radically improved many aspects of the recruiting process. But while screening has radically improved, AI can throw up new problems for hiring managers.

Studies this year by Canva and McKinsey both show candidates that use AI tools are applying for more jobs and are far more likely to land them. However those same tools are helping some people get into roles they’re not actually qualified to do. 

The Rise of the AI-Augmented Applicant

A 2025 survey by Capterra found that 73 per cent of Australian job seekers now use AI tools when applying for work, a figure that feels low given how embedded the technology has become. As Dr David Tuffley, Senior Lecturer in Applied Ethics at Griffith University, told ABC News, “AI literacy is becoming a core skill in itself. If a candidate can use AI effectively and transparently, that’s a sign of adaptability, not dishonesty.”

The operative word there is transparently. The same Capterra study revealed that nine in ten job seekers who used AI admitted to enhancing or fabricating parts of their applications, from rewriting résumés to completing test assignments. In a separate survey, the Australian Computer Society (ACS) found that 27 per cent of candidates confessed to using AI tools to generate interview responses in real time, sometimes with covert teleprompter software feeding them lines.

Research by McKinsey and MIT shows that employers are increasingly encountering candidates who sail through AI-assisted interviews or technical tests but struggle to perform once hired. For small businesses without dedicated HR teams, that mismatch can mean wasted payroll, lost productivity, and in some cases, reputational damage.

And the irony is that both sides are now using the same tools. As AI is deployed for screening on the hiring side, so too are applicants using it to polish or pad their submissions, effectively cancelling each other out.

So, what counts as legitimate AI use when you’re hiring? And where do you draw the line? According to Kate Jolly, Global Head of Talent Acquisition at Employment Hero, small businesses should expect most résumés to be AI-polished, and focus instead on verifying substance. “The problem isn’t refinement; it’s when AI fabricates achievements that don’t exist.,” she says.

The Cost of a Mis-Hire

For small businesses, hiring the wrong person can have severe financial and operational consequences. Employment Hero’s 2025 Hiring Snapshot Report reveals that Australian small businesses spend around A$8,090 annually on recruitment, while medium-sized businesses spend even more – up to A$18,371 on average. Despite this investment, 60 per cent of SMEs lose new hires within their first month, and 78 per cent have experienced a new employee leaving within three months, forcing them to restart the costly hiring process.

The fallout from a mis-hire is more than just recruitment expenses. SMEs face significant productivity losses during understaffed periods, overtime costs as 56 per cent of businesses rely on existing staff working extra hours to fill gaps, and wasted training investments when new hires leave early. Business owners and managers also spend valuable time rehiring instead of focusing on growth, all of which can add up to tens of thousands of dollars per mis-hire.

Setting the Terms for Both Sides

The Australian Public Service Commission (APSC) has released clear guidance on managing AI in recruitment, urging employers to set expectations early and candidates to be upfront about how they use the technology.

Under these recommendations, applicants can use AI to polish résumés or draft cover letters, but technical, creative, or assessment tasks must be completed independently. It’s a boundary that protects the merit-based nature of recruitment while acknowledging that AI is now part of the toolkit.

Candidates are expected to disclose any AI assistance and demonstrate their own skills during interviews or tests. Misusing AI or providing false information can be grounds for disqualification or, within the public service, disciplinary action under the APS Code of Conduct.

Once a shortlist is in place, the APSC encourages employers to move beyond polished applications and test what candidates can actually do. That means:

• Asking candidates to complete short, live tasks that show their problem-solving in real time.
• Using structured behavioural interviews that draw on genuine past experiences. “Tell me about a time you managed X situation”, rather than hypothetical scenarios.
• Verifying at least one concrete claim from a résumé through reference checks, portfolios, or tangible proof of work.

These practices shift the focus back to human capability, where AI can’t fake fluency or experience. In a hiring landscape increasingly shaped by automation, that’s where the real signal lies.