We are in the middle of a breakdown in the hiring ecosystem — and it isn’t about a single technology, but about a race.
On one side, up to 74% of job seekers now use AI to write, optimize and mass-submit applications (Greenhouse, 2025). On LinkedIn alone, application volume rose 45% in a single year — roughly 11,000 applications per minute (New York Times, 2025). Tools like LazyApply send up to 750 applications a day per candidate.
In response, companies have rolled out their own algorithmic shields. The average recruiter today handles about 2.7 times as many applications as three years ago (Gem, 2025), and among the largest companies automation is near-total: 99% of Fortune 500 firms now automate their screening process (Harvard Business School, 2024).
The result is an “AI-versus-AI” stalemate. Machines talk to machines. And somewhere in between, the thing recruitment has always been about disappears: the human judgment of whether there is a real person of real value behind the output.
The three tensions defining recruitment in 2026
When we advise HR leadership teams, the same three dilemmas keep surfacing — each putting its own pressure on the process.
1. Are we choosing the best employee, or the best AI user?
When the candidate uses AI to tailor the “perfect” CV, we are no longer measuring real competencies but the ability to prompt. Two CVs can sit side by side in the inbox: one razor-sharp, flawless, full of the right keywords — written by ChatGPT — and one a little clumsy but genuine, from the candidate who actually did the work. The risk is that we call the best AI user first. At the same time, AI flattens the differences: when every application looks alike to the algorithm, differentiation vanishes and the workload on human recruiters paradoxically grows.
2. Are we filtering out the “irregular” profiles?
AI screening optimizes for the measurable rather than the valuable. That means atypical profiles with non-linear career paths risk being filtered out automatically — precisely the profiles that often hold the resilience and creativity a company needs most. Bias can be reduced through automation, but complexity risks disappearing along with it.
3. Vetting in a world where everything can be simulated.
With deepfakes, simulated video interviews and bots that can pass technical tests, the CV has effectively stopped being proof of skill. Gartner predicts that 1 in 4 candidate *profiles* worldwide will be fake by 2028 (Gartner, July 2025). This is not just an HR problem — for an industry built on sensitive customer and personal data, it is a security risk.
From output to judgment
The common answer to all three tensions is the same shift: if output is no longer a credible signal, we have to measure the level above it — the judgment behind it.
The HR function of the future shouldn’t fight AI with an even harsher filter. Volume kills quality, and a tighter filter doesn’t solve a volume problem — it amplifies it. Instead, this is about augmentation: mastering the assessment of how a person works with AI. AI takes the pattern-recognizable mass; the human takes the uncertain, irregular edge, where the cost of an error is highest and the criteria are fuzziest.
What should HR do now?
– Introduce live cases and “vibe coding.”
Test the candidate in real time with AI tools at hand. Don’t treat the use of AI as cheating — assess the judgment: What did the candidate change in the AI’s answer? Where did it fall short? The process is the signal, not the result.
– Move toward skills-first.
Put less weight on the specific diploma and more on verifiable skills. According to McKinsey, hiring based on skills is roughly five times more predictive of job performance than hiring based on education. And employers are following suit: 71% of leaders would rather hire a less experienced candidate with strong AI skills than a more experienced one without (Microsoft & LinkedIn, Work Trend Index 2024).
– Keep the human in the loop — but move the human up, not out.Let AI flag outliers *up* for review, never *down and out*. AI may surface, enrich and remove the unambiguous — but the rejection decision in any case requiring judgment must be owned by a human. No candidate should leave the process without human approval.
The real competitive advantage
In a world where AI can support knowledge and production, professional competence becomes a hygiene factor — something everyone has, and which therefore no longer differentiates. What remains is the human qualities as the only real difference: judgment, ethics, relationships, and the ability to adapt and unlearn quickly, because the tools change every quarter.
The top candidates of the future — and the HR leaders of the future — are not the ones who know the most. They are the ones who develop the fastest in concert with both people and AI.
So the question isn’t *whether* you should use AI in recruitment. It’s whether you use it to replace people — or to make your people more human.
This is exactly the conversation we have in Sylvester & co’s HR AI Network — the network for HR leaders and specialists who need to turn AI into concrete action, with dedicated tracks for Senior CHRO´s, Snr. HR Business Partners, HRIS and Talent Acquisition. More information here; HR AI Network – Sylvester & co – HR & AI Platform | HR AI Network
If you want to go deeper into how judgment, authenticity and governance become HR’s new core competencies, join us at Nordic HR AI Summit 2027 (Nordic HR AI Summit 2027 – Sylvester & co – HR & AI Platform) (7–8 January, Copenhagen) “From Insight to Impact.”*