How AI in Recruitment Is Reshaping the Hiring Process — and What IT and Privacy Pros Need to Know

How AI in Recruitment Is Quietly Rewriting the Rules for Every Stakeholder

Artificial intelligence has embedded itself into nearly every corner of the professional world, but few areas have been disrupted as fundamentally — and as quietly — as recruitment. AI in recruitment is no longer a futuristic concept confined to Silicon Valley talent teams; it is now a practical reality across industries, affecting how roles are written, advertised, screened, and filled. For IT decision-makers, privacy professionals, and policy experts who sit at the intersection of technology governance and human resources, the implications are significant — and not all of them are positive.

David Shanahan, a director at Irish IT recruitment agency IT Search (a member of the Vertical Markets Group), has observed this shift play out across both sides of the hiring table. "AI is now being used across every stage of the recruitment process," he explained in a recent analysis for Silicon Republic. "Candidates are using it to tailor CVs, create cover letters, prepare for interviews and process job applications at scale. Employers are using AI to write job descriptions, support sourcing outreach and screen and rank applications." The speed of adoption, he notes, has outpaced the development of any meaningful governance frameworks to manage it.

For technologists and compliance officers paying close attention, the picture that emerges is one of rapidly proliferating tools operating in a near-regulatory vacuum — a situation that echoes early concerns around algorithmic decision-making in financial services, credit scoring, and content moderation. The difference is that recruitment decisions directly affect people's livelihoods, making transparency and accountability not just good practice but, in many jurisdictions, a legal obligation.

The Tool Stack Driving Automated Hiring Decisions

Understanding what is actually happening at a technical level matters enormously for professionals responsible for procuring, auditing, or governing HR technology. On the employer side, AI-assisted applicant tracking systems (ATS) now commonly apply natural language processing to rank and filter incoming applications before any human reviewer sees them. Platforms from vendors including Workday, Greenhouse, and iCIMS have integrated machine learning layers that can score candidates against job descriptions, flag keyword mismatches, and prioritise outreach lists — often without hiring managers being fully aware of how the ranking logic works.

On the candidate side, large language model tools — most prominently OpenAI's ChatGPT, but also Anthropic's Claude and a growing ecosystem of specialised products — are being used to generate and customise application materials at industrial scale. A single candidate can now apply to hundreds of roles in the time it would previously have taken to craft a single tailored application. According to research by the Society for Human Resource Management (SHRM), this dynamic is already contributing to a significant increase in application volumes per role, creating what Shanahan describes as "more noise" in the pipeline rather than better signal.

More controversially, tools designed specifically for live interview assistance — including products such as Cluely, LockedIn AI, and InterviewCoder — can analyse interview questions in real time via screen-capture or audio input and surface suggested responses directly to candidates. For hiring managers, the frustration is mounting. These are not fringe tools; they are commercially available products marketed openly to job seekers, and their use is growing. The integrity of live interviews — long considered the most reliable way to assess genuine problem-solving ability — is now in question.

GDPR, the EU AI Act, and the Compliance Minefield in AI-Assisted Hiring

For European organisations — and particularly for privacy professionals and IT compliance leads — the legal landscape surrounding AI in recruitment is rapidly becoming more complex. The General Data Protection Regulation (GDPR) has always had specific implications for automated decision-making: Article 22 gives individuals the right not to be subject to decisions based solely on automated processing when those decisions produce significant legal effects. A rejected job application almost certainly qualifies.

Yet in practice, enforcement has been inconsistent, and many organisations deploying AI screening tools have not fully mapped their processing activities to their GDPR obligations. This is not a theoretical risk. The Irish Data Protection Commission and other EU supervisory authorities have signalled increasing interest in HR data processing, and several formal investigations have already been launched against organisations using AI profiling tools without adequate transparency notices or data protection impact assessments (DPIAs).

The EU AI Act, which entered into force and is being phased in progressively, adds another layer. AI systems used in employment contexts — including CV screening, candidate ranking, and interview analysis — are classified as high-risk systems under the Act. This designation triggers mandatory requirements around transparency, human oversight, accuracy testing, and bias auditing. Organisations that procure these tools from third-party vendors cannot simply outsource accountability; under both the AI Act and GDPR, the data controller remains responsible for ensuring that processing is lawful and fair.

For IT decision-makers evaluating HR technology vendors, this creates a clear due diligence checklist. Vendors should be able to demonstrate how their models were trained, whether bias audits have been conducted and on what demographic datasets, how explainability is implemented, and what data retention policies apply to candidate information. According to guidance published by the European Data Protection Board, candidate transparency notices must specifically disclose any automated processing and the logic involved — a requirement many current deployments fall short of meeting.

Separating Genuine Capability from AI-Polished Output — A Practical Framework

The core challenge that Shanahan identifies is not whether AI is being used, but whether its use obscures what employers actually need to know: can this person do the job? It is a question that has significant resonance for technical hiring in particular. When recruiting software engineers, security analysts, DevOps professionals, or data architects, a polished AI-generated cover letter reveals precisely nothing about whether the candidate can debug production code under pressure or architect a zero-trust network.

The practical countermeasures Shanahan recommends are grounded in a simple principle: design assessments that cannot be meaningfully gamed by AI, rather than attempting to prohibit AI use outright. Specifically, this means moving away from questions with single correct answers — the type that LLMs handle trivially — and toward real-time problem-solving exercises that require candidates to explain their thinking, adapt to new information mid-process, and justify decisions in dialogue. A candidate who genuinely understands distributed systems architecture can explain trade-offs on the fly; one who has used AI to memorise a canned answer typically cannot.

On the technical integrity side, Shanahan also recommends that employers concerned about live AI assistance during remote interviews ask candidates to share their full desktop — not just a single browser tab or application window — since many AI interview tools operate as browser extensions, overlay applications, or separate desktop processes. While this is not a foolproof solution, it raises the bar meaningfully. Some organisations are also exploring AI detection tooling specifically designed for interview environments, though the accuracy and fairness of such tools remain contested.

Why Transparency Is the Non-Negotiable Foundation of Fair AI Hiring

Across both the regulatory and ethical dimensions of this debate, one principle consistently emerges as foundational: transparency. Shanahan is explicit on this point: "If employers are using AI for candidate screening, assessments or interview tools, candidates should be informed. Equally, candidates should feel comfortable raising concerns if they believe AI is having a negative impact on the process." This is not merely a best-practice recommendation — under GDPR's transparency and fairness principles, and under the forthcoming obligations of the EU AI Act for high-risk systems, it is increasingly a legal requirement.

For organisations looking to establish credible AI governance in their hiring processes, this means going beyond a one-line disclosure in a privacy policy. It means implementing a clear AI transparency policy that addresses: which stages of the process involve automated tools; what data is being processed and for how long; how candidates can request human review of automated decisions; and what recourse exists if a candidate believes AI has produced an unfair outcome. Research from McKinsey's talent analytics practice suggests that organisations with clearly communicated AI policies in their hiring processes report higher candidate trust scores and lower dropout rates — a commercial benefit that aligns with the compliance imperative.

The transparency obligation cuts both ways. As AI tools become standard in candidate preparation, employers who silently use AI screening while penalising candidates for AI-assisted applications are operating in bad faith. The more defensible and sustainable position is one in which both parties are open about the tools they are using, and where the human judgement that AI is supposed to support remains genuinely visible in the outcome.