Applying for a job today requires a very different skill set than it did even five years ago. Increasingly, applicants use AI to tailor resumes in ways they hope will pass automated screening.
While AI-powered recruitment tools promise efficiency and reduced bias, their widespread use has unintentionally gamified hiring and may deepen a digital divide among job seekers.
Applicant tracking systems (ATS) have long been used to screen and identify strong applicants. Before AI-driven screening, ATS platforms largely relied on keyword matching and screening questions to filter and rank candidates for recruiter review and shortlisting.
AI-enhanced ATS platforms now parse resumes using natural language processing (NLP), converting unstructured text into structured data for analysis. They score and rank applicants against job requirements, with the expectation of producing a more objective match for recruiters to review.
Observation
“Applicants are either answering the questions incorrectly to get by the automated screening, or they are lying about their experience.”
As AI screening becomes more common, it raises questions about tool design and unintended consequences.
The application process has become gamified – transforming hiring into a data-driven game where success depends less on qualification and more on technological fluency. Recruiters use AI powered screening tools to score and rank applicants. While applicants use them to tailor resumes meant to get past the AI screening tools.
Some recruiters now report increased workloads, spending more time reviewing resumes and conducting additional interviews, as applicants use AI to strengthen, even overstate, their skills and experience.1
In a brief chat with Google Gemini, it gave me tips to pass AI screening, including suggested keywords, a resume structure, and ways to quantify achievements. Gemini also instructed me how to prepare for an AI-powered video screening interview, noting that it may assess indicators like eye contact, facial expressions, and confidence (I am particularly interested in how these variables are defined and analyzed). These instructions help to normalize gaming behaviours.
Many ATS vendors position resume screening as a way to reduce bias through standardized criteria. This claim warrants careful scrutiny in the design, deployment, and ongoing monitoring of these tools.2 A recent study shows how LLMs exhibit “self-preference bias” when reviewing resumes, preferring resumes written by an LLM, or more troubling, the same model the screening tool was built with.3
At the same time, screening tools marketed as bias-reducing can reinforce digital inequality. Applicants with access to AI tools and the skills to use them may be at an advantage, while job seekers who lack these skills, lack the resources to become AI literate, or have limited access to a computer and AI tools may be excluded. The introduction of new technology can reinforce and deepen existing inequalities between those who are AI literate and make it harder for those disadvantaged.4
Transforming Recruitment Norms and Practices
Many organizations have not fully considered the implications of using AI in recruitment.
It appears organizations and applicants have unknowingly contributed to a new dilemma fueled by artificial intelligence: reinforcing digital inequality. Automated screening encourages applicants to optimize (and sometimes inflate) resumes with AI, which in turn drives demand for more automation to manage volume and variability. The growing reliance on AI-driven applicant screening illustrates a systems-level failure in recruitment design: where incentives meant to improve efficiency instead reinforce resume inflation, human resource workarounds, and unequal access to opportunity.
How can organizations respond? A systems lens and human-centred design approach are a strong starting point. Key questions include:
- Problem Definition: What problem is being solved? Which parts of recruitment require human judgement – a human-in-the-loop? Where can automation best support recruiters so they can focus on assessing applicant capability and fit?
- Consider unintended consequences: Has the ATS vendor considered how their application will correct for algorithmic bias? What mechanisms are in place (human or otherwise) that will assess whether the automated process yields a representative candidate pool?
- Invest in AI Adoption: What investments are being made to support recruiters and hiring leaders to use AI-powered recruitment tools? Undertanding LLMs enables organizations to question and design purpose-driven tools rather than implement out-of-the-box-one-size-fits-all solutions (they don’t).
At a national level, we can collectively support publicly funded AI literacy education, including for older adults and newcomers, and improve access to AI tools and training across sectors.
The introduction of automated resume and applicant screening tools are helpful but risk reducing individuals to structured data and turning hiring into a data game. My challenge to screening-product designers, organizations, and recruiters is to build technology and design processes so all applicants can demonstrate their protentional on equal terms.
AI is changing the recruitment game, and organizations can play a role in responsible design and adoption.
- https://www.hcamag.com/ca/specialization/transformation/ai-tools-slow-hiring-add-pressure-on-hr-teams/568007 ↩︎
- https://www.hr.com/en/magazines/all_articles/workday%E2%80%99s-ai-bias-lawsuit-grows%E2%80%94a-cautionary-tale-_mcltvlyd.html ↩︎
- https://arxiv.org/abs/2509.00462 ↩︎
- https://www.forbes.com/sites/hessiejones/2025/02/26/politics-and-the-perils-of-ai-exacerbating-social-divides-in-canada/ ↩︎