Analysis: Job hunters have always found ways to push back against HR selection systems and today's AI-led recruitment process is no different
Employers have always searched for a recruitment and selection process that reliably identifies the right person for the job, with varying degrees of success. In the early 1900s, one Philadelphia manufacturer selected workers by tossing apples into a crowd: you were hired if you caught one and kept it. By the 1950s, corporations had swung to the opposite extreme, subjecting candidates to days of IQ tests and personality assessments.
The mid-2000s saw a Google-inspired fad for brainteaser questions ("How many golf balls fit in a school bus?") that was quietly dropped once research showed it bore no relationship to on-the-job performance. Even structured interviews, widely regarded as the most reliable method of selection, predict only about half the variation in job performance.
Now, AI promises to improve on that track record, According to a recent World Economic Forum report, 90% of employers use some form of automated system to prioritise, rank or deselect candidates. Historically, candidates have found ways to push back against whatever system is popular at the time. When psychometric tests became common, a niche coaching industry sprang up around them. When competency-based interviews took hold, candidates learned to structure every answer around a standard format.
We need your consent to load this rte-player contentWe use rte-player to manage extra content that can set cookies on your device and collect data about your activity. Please review their details and accept them to load the content.Manage Preferences
From RTÉ Radio 1's The Business, is the AI bubble about to pop?
Today's AI-led recruitment is no different, except that both sides now have access to the same technology. Candidates use generative AI to mass-produce tailored applications; employers respond with ever more sophisticated screening tools. For most job-seekers, these tools are now the first gatekeeper to getting access to employment. If your application does not get past the screening algorithm, no human will ever read it.
However, this technology is far less sophisticated than many candidates assume. In Hilke Schellmann's excellent book The Algorithm, she tested many of these screening processes herself. In one case, she answered every question in a one-way video interview by repeating "I love teamwork" and received a high score. In another, she responded entirely in German to questions for an English-language role and was rated 73% qualified. If a system cannot distinguish a genuine answer from a nonsensical one, candidates should feel less intimidated by it and more strategic about how they engage with it.
How do you get past the screening process?
Getting through the screening process involves fairly mundane things. Screening algorithms are focused on matching keywords from the job description against your CV, so read the posting carefully and make sure the language lines up. Do not just copy the whole description in, as some systems will penalise that, but aim for roughly 80 to 90% overlap. Spell out any acronyms alongside any abbreviations you use too. If the posting says "Search Engine Optimisation" and you have written "SEO", the system may not make that connection.
We need your consent to load this rte-player contentWe use rte-player to manage extra content that can set cookies on your device and collect data about your activity. Please review their details and accept them to load the content.Manage Preferences
From RTÉ Radio 1's Oliver Callan, should you use AI to write a CV?
It is also worth including a dedicated skills section, and not just a generic list of competencies. A lot of companies are moving toward skills-based hiring, so break out your technical skills (software, languages, certifications) separately from your interpersonal strengths (stakeholder management, team leadership, conflict resolution). Do not bury them three bullet points deep into a job description from 2019.
Format is the part that catches a lot of people off guard. Most applicant tracking systems read your document as a simple stream of text. If you use a two-column layout, or put your contact details in a header, or include graphics and text boxes, the system may treat those elements as objects rather than text and skip over them entirely. There is some evidence that roughly one in four rejections happens because the system simply could not read the file.
Specificity is also important. Vague claims like "improved efficiency" mean very little to an algorithm or a human, whereas "Cut processing time by 30%" gives both something to work with.
Caution needs to be expressed over the oft-repeated "white fonting" hack, which goes viral on TikTok every year or so. This involves pasting the job description in white text on your CV so it is invisible to the human eye but readable by the AI system. ManpowerGroup, the third largest staffing firm in the world, reports detecting hidden text in around 100,000 applications per year that they receive, and candidates who are caught doing this are not moved forward.
A more practical point that is often overlooked is to apply directly through company websites rather than job boards or recruitment platforms. Those working in HR consistently report checking their own system’s submissions first.
Navigating AI interviews and assessments
The application stage is only the beginning. Employers increasingly deploy AI across the entire hiring process through structured AI-led interviews that score your verbal responses, chatbot screening calls, game-based assessments, and video analysis tools that purport to evaluate communication skills. There is a telling irony here.
We need your consent to load this rte-player contentWe use rte-player to manage extra content that can set cookies on your device and collect data about your activity. Please review their details and accept them to load the content.Manage Preferences
From RTÉ Radio 1's The Business, recruitment agent Louise Campbell on what's changed with job interviews and how much of this is the fault of AI.
Having steadily removed human beings from hiring, many companies now try to make their AI feel more human. Chipotle’s "Ava Cado" chatbot will talk you through their hiring process, and is one example of the impulse to anthropomorphise what are simply pattern-matching engines. All of these tools supposedly convert your responses into predictive indicators of future performance.
For candidates, this means the hiring process is about the signals you produce across multiple touchpoints. In AI-led video interviews, speak clearly and directly to the camera. Use short, structured answers focused on concrete examples and specific outcomes rather than warmth or charm.
For chatbot screening, answer the actual question asked and avoid volunteering tangential information; these systems generally score against a very narrow rubric. Game-based or psychometric assessments offer less room to prepare, but knowing they typically measure problem-solving speed or pattern recognition can help you approach them with less anxiety.
We need your consent to load this rte-player contentWe use rte-player to manage extra content that can set cookies on your device and collect data about your activity. Please review their details and accept them to load the content.Manage Preferences
From RTÉ Radio 1's The Business, will your job be replaced by AI?
Use AI wisely, but sound like yourself
Candidates are already using AI to write applications, and this is a rational response to AI-led screening. If employers can use AI to filter you out, is there anything wrong with using AI as an aid to give yourself a better chance of getting through? Many people think of it as levelling the playing field and will paste the job description alongside their CV into their LLM of choice and ask it to identify gaps in keyword alignment. They may also use AI to research the employer's competitors, recent news, and salary benchmarks. That kind of contextual knowledge is exactly what generic applications lack.
But follow this logic through. If employers and applicants both continue to outsource their thinking to AI, the endpoint is a dystopian farce with personal AI agents writing and submitting job applications on our behalf, the employer’s AI screens and scores them, an AI interviewer conducts the first round interview, and a job that goes to whichever candidate has the better algorithmic score.
We are not there yet, thankfully, and there is some room for optimism. If every candidate uses the same AI tools, every application starts to read the same way, and the written material stops being a useful way to tell people apart.
When the written application loses its value, the parts of the process where you come across as an actual person start to carry a lot more weight.
An interviewer who realises you actually know what the organisation does, or a hiring manager who reads a cover letter and thinks "this person gets it", will always matter more than whatever score an algorithm generated. No tool can fake that at scale. Use AI to sharpen your application, by all means. But when a human finally reads what you have written, make sure they find an actual person on the other side of it.
Follow RTÉ Brainstorm on WhatsApp and Instagram for more stories and updates
The views expressed here are those of the author and do not represent or reflect the views of RTÉ
