AI can support recruitment decisions, but human judgement must still lead the final hiring choice. Photo by Canva
AI Is Changing Recruitment — But Who Should Make the Final Hiring Decision?
Artificial intelligence is quickly becoming part of recruitment.
It can help recruiters process large volumes of applications, organise candidate information, identify relevant skills, automate administrative work and make the hiring process faster.
But there is a bigger question employers now need to answer:
Just because AI can help make a hiring decision, should it make the decision?
That question has become especially relevant following Singapore’s latest push towards a people-centred AI transformation for HR.
In September 2026, the Tripartite Workgroup on Human Capital Capability Development recommended developing guidance for the responsible use of AI in recruitment, including processes and safeguards to ensure employment decisions remain fair and merit-based. The Government has accepted the workgroup’s recommendations.
The message for employers is not to stop using AI.
It is to use AI without removing human judgement from hiring.
The message for employers is not to stop using AI.
It is to use AI without removing human judgement from hiring.
Why Responsible AI in Recruitment Matters Now
For many employers, recruitment is one of the easiest places to see the benefits of AI.
Imagine a company receives 1,000 applications for one vacancy.
Instead of a recruiter manually opening every CV, an AI-enabled recruitment system could help organise applications, extract relevant experience, identify skills and surface candidates who appear to match the role.
That can save significant time.
But now imagine that the system’s definition of a “good candidate” is too narrow.
A candidate could have transferable skills but a different job title.
A career switcher may not have the conventional experience normally associated with the role.
A returning worker may have an employment gap.
An older candidate may have decades of useful experience but a CV structured differently from younger applicants.
A candidate may simply use different keywords from those the system expects.
The technology may be working exactly as designed — while the employer still misses potentially suitable talent.
That is why responsible AI in recruitment is not simply an IT issue.
It is a hiring strategy issue.
AI Can Screen Information. Humans Still Need to Understand People.
The most useful way to think about AI recruitment is not:
AI versus recruiters.
It is:
What should AI do, and what should humans continue to decide?
AI is particularly useful when dealing with repetitive and high-volume recruitment tasks.
It can help organise candidate information, compare applications against defined criteria, support scheduling and reduce administrative work.
Human recruiters, however, can consider things that may not be obvious from structured data alone.
They can ask:
- Does this person’s experience transfer to this role?
- Is there a reasonable explanation for the career gap?
- Does the candidate have potential that isn’t obvious from the job titles on the CV?
- Could someone from another industry perform this role successfully?
- Are we rejecting candidates because they genuinely lack the capabilities required — or simply because they don’t resemble the people we traditionally hire?
This distinction becomes increasingly important as companies move towards skills-based hiring, career conversion, senior workforce participation and job redesign.
Singapore’s fair employment framework already expects employers to recruit and select based on merit, including skills, experience and the ability to perform the job, rather than non-job-related characteristics.
The Risk Is Not Always the AI. Sometimes, It Is the Criteria We Give It.
There is sometimes an assumption that removing humans automatically removes bias.
It isn’t that simple.
An AI recruitment tool ultimately operates using the information, rules, data and objectives provided to it.
For example, suppose an employer tells a system:
Prioritise candidates who have held the exact same job title for at least five years.
The system may perform that instruction perfectly.
But the hiring criteria itself could eliminate candidates with highly relevant transferable skills.
Likewise, if businesses automate an outdated recruitment process without first reviewing it, AI may simply allow that old process to operate faster and at greater scale.
This is why responsible AI starts before the technology is switched on.
Employers first need to ask:
What actually makes someone capable of doing this job?
That may mean moving beyond qualifications and past titles towards skills, capabilities, experience and potential.
What Human-Led AI Recruitment Actually Looks Like
Human-led recruitment does not mean recruiters need to manually perform everything.
The objective should be to let technology handle appropriate tasks while retaining human accountability over consequential employment decisions.
A practical model could look like this:
AI can support:
- Organising applications
- Extracting skills and experience
- Identifying potential matches
- Handling repetitive administration
- Supporting interview scheduling
- Providing recruitment insights
Humans should remain accountable for:
- Defining what makes a candidate suitable
- Reviewing whether criteria are genuinely job-related
- Considering transferable skills and non-traditional backgrounds
- Investigating unusual or borderline cases
- Conducting meaningful candidate conversations
- Making and taking responsibility for hiring decisions
- The point isn’t that every AI recommendation must be rejected or manually recreated.
It is that employers should be able to understand why candidates are being progressed or rejected and intervene when necessary.
Singapore’s Ministry of Manpower has similarly stated that employers remain responsible for complying with fair and merit-based employment practices even when AI tools are involved in employment decisions.
5 Principles for Responsible AI Recruitment in Singapore
1. Start With Job-Relevant Criteria
Before automating candidate screening, review the criteria being used.
Separate genuine requirements from historical preferences.
For example:
Necessary: ability to analyse financial data.
Potentially unnecessarily restrictive: must have held exactly the same job title previously.
The more accurately the company understands the capabilities required for the job, the more useful AI becomes.
2. Keep a Human Checkpoint
AI can recommend.
Humans should remain accountable.
This becomes especially important when candidates fall outside conventional recruitment patterns.
An experienced recruiter may recognise potential in someone whose career history does not perfectly match the vacancy but whose skills do.
This is where human judgement adds value rather than simply duplicating what the technology has already done.
3. Review Outcomes, Not Just Efficiency
A recruitment system shouldn’t be measured only by:
How much faster did we shortlist candidates?
Employers should also examine:
Who is being screened out?
Are suitable candidates repeatedly being missed?
Are our hiring criteria unnecessarily narrowing the talent pool?
Are recruiters routinely overriding certain recommendations?
These questions can reveal whether the technology is genuinely improving recruitment or simply accelerating the existing process.
4. Make Employment Decisions Explainable
As AI plays a larger role in decisions affecting people’s careers, trust becomes increasingly important.
MOM has highlighted the importance of explainability, feedback mechanisms and human oversight in responsible workplace AI adoption.
Employers therefore need to understand how their recruitment technology is being used.
If nobody in HR can explain why a system continually recommends or excludes particular candidates, that should be a warning sign.
5. Redesign Recruitment Before Automating It
This may be the most important principle.
Do not automate a recruitment process simply because it already exists.
Review it first.
Ask:
- Which recruitment steps create real value?
- Which are repetitive?
- Where are good candidates dropping out?
- Are job requirements still relevant?
- Which decisions require judgement?
- Which activities can safely be supported by AI?
This reflects the wider direction of Singapore’s latest people-centred AI recommendations: technology decisions and workforce decisions should be considered together rather than separately. The workgroup specifically highlighted job redesign and the need to consider jobs and skills while transformation decisions are being made.
Fair Hiring Is Becoming More Important, Not Less
The conversation is particularly timely because Singapore’s employment framework is also evolving.
The Workplace Fairness Act is currently targeted to take effect in end-2027. It strengthens protection against adverse employment decisions based on protected characteristics including age, nationality, sex and related characteristics, race and religion, disability and mental health conditions.
AI does not remove an employer’s responsibility for those decisions.
MOM has stated that employers must comply with fair and merit-based employment practices when deploying AI tools for hiring and other employment decisions.
For employers, this creates an important principle:
Technology may support the hiring process. Accountability still sits with the employer.
The Future Recruiter Is Not Being Replaced — the Job Is Changing
As AI takes over more administrative recruitment work, the role of the recruiter should evolve as well.
Less time may need to be spent sorting information.
More time can potentially be spent understanding candidates.
That means capabilities such as judgement, contextual understanding, communication, persuasion, relationship-building and the ability to recognise transferable potential become even more important.
The recruiter of the future should not compete with AI at processing CVs.
The recruiter should provide what the technology cannot fully provide:
human context behind the data.
That is what people-centred AI in recruitment can look like in practice.
How Elitez Approaches the Future of Recruitment
As a recruitment and workforce solutions provider in Singapore, Elitez supports employers across permanent recruitment, temporary and contract staffing, executive search, workforce transformation and HR advisory.
Our view of AI adoption is straightforward:
The objective should not simply be to automate more recruitment activities.
It should be to build a better recruitment process.
That means examining the job, the skills required, the workflow, the technology and the human decision points together.
The same principle applies beyond recruitment. Through workforce transformation and Job Redesign+, businesses can review existing roles and processes before introducing technology, rather than placing new AI tools on top of inefficient workflows.
AI can make recruitment faster.
Human judgement helps make sure faster still means better.
And as Singapore moves towards more people-centred AI adoption, employers that combine both will be better positioned to build hiring processes that are efficient, fair and capable of recognising talent that others may overlook.
Is Your Hiring Process Ready for AI?
Before introducing AI into recruitment, understand which parts of your hiring process should be automated, redesigned or remain human-led.
Elitez helps organisations review recruitment, workforce and job design to build more productive, skills-focused and future-ready teams.