Be Careful What You Wish For
On 18 August 2025 the Financial Times asked, “Does HR still need humans?” It reported how AI is transforming HR functions, from handling employee queries to filtering job applications. Efficiency gains are clear, but the article underlined the questions of bias, judgment, and trust that follow when people processes are handed to machines.
At the same time, the largest HR technology vendors are selling a seductive vision. Workday, Oracle, SAP SuccessFactors and UKG all promise “AI powered” platforms that will cut costs, accelerate recruitment, and improve the employee experience. Consulting firms amplify the message, pitching these tools as the cure for everything that ails HR.
We argue that the promised efficiencies risk creating an AI-to-AI recruitment loop, in which neither the applicant nor the employer plays a meaningful role in the first stages of hiring. The result is not progress but a parody of it.
The pitch from the platforms
- Workday highlights its new “AI agents” that can automate HR queries and accelerate hiring, with hundreds of AI enhancements built into the platform.
- Oracle markets recruiting tools that “streamline the hiring process through AI and automation,” alongside a suite of AI enabled HCM features.
- SAP SuccessFactors promotes AI assisted writing, text analysis and skills extraction across job profiles, CVs and role descriptions.
- UKG introduces “Bryte” AI agents to simplify people processes and improve candidate experiences.
The marketing is consistent. These systems promise to save time, reduce cost and scale recruitment without sacrificing quality.
We say: be careful what you wish for.
The displacement of human agency
What happens when both sides of recruitment are automated? Employers use AI to screen CVs and covering letters. Applicants use AI to draft them. The first exchange is no longer human at all. One machine scores the output of another.
This is part of a broader trend, the gradual displacement of human agency. We are seeing it in law, finance, education, and most other sectors. The individual cedes control to the system, and the system itself is run by AI. In recruitment, this means the applicant no longer represents themselves and the employer no longer evaluates directly. What remains is a machine-mediated proxy for both sides.
Recruitment should be about matching talent with opportunity. It should allow space for originality, judgment, and human potential to come through. Instead, the process risks being reduced to algorithmic pattern matching between AI generated content and AI powered filters.
The corporate incentive problem
Why does this happen? The answer lies often in corporate incentives. For most large organisations, HR is viewed primarily as a cost centre. AI offers dramatic savings by reducing the number of people involved in screening, shortlisting, and administration. As long as the process produces “acceptable” hires, there is little pressure to revisit it.
This logic does not apply evenly. For mission-critical roles, organisations will still invest in rigorous, human-led recruitment. But for the majority of roles, the priority will remain efficiency over talent. The absurdity becomes tolerated because it is cheap.
The false comfort of live testing
Some organisations argue that the problem can be solved with live skills assessments. Coding challenges, problem solving exercises, and collaboration tasks are used to cut through the AI haze. At first glance this looks like progress.
In reality, it does not resolve the deeper problem. Only those who pass the initial automated screen reach the live stage. If strong candidates are filtered out early, the quality of the pool is already compromised. And when most assessments are conducted online, applicants can and will use real time AI support to pass them. Employers are drawn into an arms race of monitoring and proctoring that quickly collides with privacy concerns.
Why this may not hold forever
There are limits to this trend. Companies that overlook good candidates damage their brand and weaken performance. Regulators are already scrutinising algorithmic bias and will push for greater transparency and human oversight. In knowledge-intensive sectors the cost of a poor hire is too high to leave recruitment entirely to machines.
There are also new models emerging. Some firms are moving away from the traditional CV and covering letter entirely, focusing instead on portfolios, structured interviews or project-based auditions. These formats are harder to fake with AI (for now), although not immune.
A test for leaders
Before investing in “AI powered” HR solutions, leaders should ask:
- How many applicants are rejected solely by automated filters, and is there human review for potential false negatives?
- How do we detect AI written applications, and how do we assess originality?
- Do our assessments measure skills that matter, or only skills that AI can replicate?
- How do we safeguard online tests without invading privacy?
- Can we explain, in plain English, why one candidate is preferred over another?
- Are we tracking the performance of hires sourced through AI screening against those chosen with more human involvement?
- Are we willing to pay more for a smaller but stronger talent pool?
Conclusion: efficiency or absurdity
The FT posed the question of whether HR still needs humans. The answer should be obvious. Of course it does. But the more important point is that humans need to remain in HR not only to manage people, but to protect the integrity of recruitment itself.
Left unchecked, AI in HR risks becoming a parody of efficiency. It creates the illusion of rational, data driven decision making, while in reality producing a machine-to-machine loop that sidelines the very talent companies are supposed to be seeking.
This is not simply an HR issue. It is another case study in the gradual displacement of human agency. If we allow recruitment to drift into automation without challenge, we risk undermining not just hiring but the very principle that work should connect people with opportunity.