In HR artificial intelligence has quickly moved from experimentation to being embedded in infrastructure. Recruiters use it to rank candidates, assessment providers use it to score responses, HR platforms use it to recommend applicants, flag risks, and automate decisions that once sat entirely with people.
Many organisations adopted these tools quickly because the pressure on HR teams to hire more efficiently is constantly growing. Recently, recruiters have faced; larger applicant volumes, skills shortages, remote hiring challenges, the dreaded pandemic, ever-growing expectations around speed and now AI.
At the beginning AI appeared to offer relief. It promised faster screening, lower administration, and more consistency across hiring decisions. Yet many HR leaders now face a difficult reality. The technology moved into recruitment faster than governance did.
Our latest research report found that 94% of HR and talent professionals use talent assessments in some form, while 51% confirmed that AI already plays a role in how those tools operate. At the same time, only 22% described themselves as “very confident” that AI is being used ethically and responsibly.
A recruiter can no longer rely on vendor claims alone. A hiring manager cannot afford to treat algorithmic decisions as neutral simply because software produced them. An HR director cannot assume that fairness exists because a platform claims to reduce bias.
When AI influences hiring outcomes, transparency and audit stop being optional safeguards. They become operational requirements.
Why HR teams are cautious
Many HR professionals no longer question whether AI belongs in recruitment. They question whether they can properly understand and defend the systems they already use.
The concern is no longer theoretical.
In our research we mention the use of “Shadow AI”. AI that is used without fully understanding how, when or where.
Around one-third of HR professionals we surveyed either suspected AI was being used in their assessment tools without confirmation or admitted they simply did not know. That means many organisations are allowing algorithms to influence hiring decisions without clear visibility into how those systems function.
For HR leaders, that creates several problems simultaneously.
A recruiter may struggle to explain why one candidate progressed while another did not. A candidate may challenge a hiring decision. A regulator may ask how an organisation tested for bias. A hiring manager may assume the system is objective without understanding its limitations.
The issue becomes larger when people inside the organisation trust outputs they cannot interrogate.
This is why transparency around these tools is so critical.
A psychometric test provider should be able to explain, in plain English, how an assessment works, what the AI measures, how scoring operates, and how bias monitoring takes place. HR teams should not need advanced technical expertise to understand the foundations of a system influencing employment decisions.
When a provider hides behind phrases such as “proprietary algorithm”, the organisation buying the tool absorbs much of the risk.
Why is AI important for HR?
AI matters because modern recruitment has become difficult to manage at scale with outusing some form of automation.
A recruiter may receive hundreds or thousands of applications for a single role. A talent acquisition team may need to assess technical skills, behavioural traits, communication ability, and role fit across multiple regions simultaneously. Human-only processes struggle under that pressure.
AI can help HR teams in several practical ways:
- Ranking candidates against job criteria
- Supporting skills-based hiring
- Reducing repetitive administration
- Identifying patterns across large applicant pools
- Accelerating shortlisting
- Supporting structured assessment scoring
- Improving consistency across hiring stages
Many organisations also hope AI can help to reduce human bias by introducing more structured decision-making into recruitment. That ambition partly explains why adoption accelerated so rapidly.
Yet AI only becomes valuable when the underlying assessment process is scientifically grounded and carefully monitored. A flawed process executed faster simply scales the problem.
The strongest organisations increasingly ask difficult questions before implementation:
- Who validated this tool?
- How often is bias reviewed?
- Can candidates understand how they were assessed?
- Is there meaningful human oversight?
- What happens if the system produces unfair outcomes?
- Can recruiters override recommendations?
- How does the organisation defend decisions legally and ethically?
These questions signal a more mature phase of AI adoption.
The growing pressure around fairness
Fairness now sits at the centre of the AI conversation in HR.
The concern appears repeatedly across recruitment teams because employment decisions carry real human consequences. A scoring model does not merely process information. It influences whether a person gets shortlisted, interviewed, promoted, or rejected.
That responsibility changes the standard expected from HR technology.
Our research found that unfair or inaccurate evaluations represented one of the largest concerns among HR professionals. Respondents also expressed concern about legal exposure and lack of transparency. These fears are interconnected because an unfair process creates legal vulnerability.
Some organisations still treat fairness as a procurement exercise rather than an operational responsibility, while others expect any due diligence will already be done by the vendor.
A buyer may ask a vendor about bias during implementation, then never revisit the issue again. An HR team may assume that compliance exists because the supplier says so. Meanwhile the assessment continues processing candidates without ongoing audit or independent review.
That creates what the report describes as an “audit gap”.
The gap matters because AI systems can drift over time. Hiring patterns change. Candidate behaviour changes. Job requirements evolve. A system that appeared fair during implementation may produce different outcomes later.
Continuous monitoring therefore becomes essential.
Independent audits, adverse impact reviews, technical validation, and regular oversight should become standard practice rather than exceptional practice.
The organisations taking this seriously are likely to gain a competitive advantage because candidate trust increasingly influences employer reputation.
The cheating arms race
Another pressure shaping HR strategy involves candidate use of AI.
Generative AI has changed how candidates approach recruitment. Applicants now use language models to write CVs, improve written responses, prepare interview answers, and potentially manipulate online assessments.
The research found that 26% of respondents had already seen evidence of candidates using AI to manipulate results, while another 36% believed it was likely happening inside their organisation.
This creates tension for some recruiters.
A hiring manager wants efficiency but also needs confidence that assessment results reflect genuine capability rather than prompt engineering skills.
Different assessment formats carry different levels of risk.
Written application questions remain highly vulnerable because AI excels at generating polished text. Traditional CV screening also becomes weaker when candidates can produce near-perfect application materials within minutes. Behavioural assessments, bespoke situational assessments, timed problem-solving tasks, and scientifically validated psychometric tests tend to be more resistant because they measure qualities harder for AI to replicate convincingly.
This shift is pushing many organisations toward skills-based hiring and behaviourally grounded assessment methods. Recruiters increasingly want evidence of real capability rather than polished presentation.
What does the future of AI and HR look like?
The future of AI in HR will likely become more structured, regulated, and evidence-driven.
The early phase of AI adoption focused heavily on speed. Organisations wanted automation quickly because recruitment pressure intensified. Many vendors marketed efficiency gains aggressively. Buyers often prioritised functionality over governance. That phase appears to be ending.
HR leaders now show stronger interest in explainability, auditability, and candidate trust.
The future will probably include several major shifts.
Greater transparency requirements
Candidates will increasingly expect employers to explain how AI contributes to hiring decisions. Regulators are also moving in this direction.
Under the EU AI Act, employment-related AI systems fall into a high-risk category requiring governance, documentation, monitoring, and human oversight.
This means organisations will need clearer policies, stronger records, and defensible assessment processes.
More skills-based assessment
Traditional CV screening is losing credibility because generative AI makes it easier to fabricate polished applications.
Recruiters will likely rely more heavily on skills based assessments measuring practical skills, behavioural traits, judgement, and applied capability.
Human oversight becoming mandatory
Most HR professionals do not want fully autonomous hiring systems. They want AI to support decision-making rather than replace it entirely.
AI may rank, score, organise, and recommend but a human still needs accountability for the final outcome. The organisations performing best in this environment will probably treat AI as a co-pilot rather than an authority figure.
Candidate experience becoming strategic
Heavy-handed surveillance may damage trust and discourage strong applicants.
The strongest employers will likely focus on creating assessment experiences that feel relevant, respectful, transparent, and difficult to game naturally through design rather than excessive monitoring.
Candidate experience and assessment integrity increasingly reinforce each other.
The new responsibility facing HR leaders
AI has altered recruitment operations permanently.
The question facing HR leaders is no longer whether AI should exist inside hiring. In many organisations it already shapes recruitment at multiple stages.
The more important question concerns accountability.
A recruiter should be able to explain how a candidate was evaluated. An HR director should understand where AI influences decisions. A vendor should provide independent evidence supporting fairness claims. A hiring process should withstand scrutiny from candidates, regulators, and leadership teams alike.
All of that requires both transparency and audits.
And it requires organisations willing to treat governance around these tools seriously.
The companies building trust around AI today are unlikely to be the ones with the loudest marketing claims. They will be the organisations prepared to show their work, validate their systems independently, and keep people genuinely accountable throughout the hiring process.
