Let’s give credit where it’s due: Bending Spoons is a remarkable company. In a few short years, it has built a world-class portfolio of digital products and reached a valuation that would make many big tech firms envious. Its operational discipline, high-performance culture, and ability to attract global talent are, quite simply, impressive.
But last week the Wall Street Journal revealed something about their hiring process that deserves a closer — and more critical — look.
800,000 applicants a year. An acceptance rate below 0.05%. Harder to get into than Harvard, they say. Headlines write themselves. Candidates talk about it on Reddit. LinkedIn lights up. And Bending Spoons, without spending a cent on employer branding, becomes the most coveted workplace in tech.
Convenient, isn’t it?
Because before we ask whether this process is legal or ethical, we should ask a more basic question: is this genuinely a hiring methodology, or is it primarily a marketing strategy or an employee retention tool? The mystique of extreme selectivity turns the company into a rare pearl — an exclusive club where only ”the chosen” are admitted. It triggers exactly the psychology that luxury brands have exploited for decades: the harder something is to obtain, the more desirable it becomes. It is a masterclass in brand building. Whether it is good HR is a very different question.
So let’s talk about the process itself.
An algorithm — trained on the historical performance data of existing employees — evaluates, scores, and eliminates candidates at scale. Reasoning tests, yes, but also behavioral and personality traits: «politeness,» for instance, is an actual weighted variable in the model. And the system feeds continuously on internal performance data in a closed loop.
This is where the uncomfortable questions begin. Is this legal under European law? (Bending Spoons is an Italian company) And even if it is, is it ethical?
The GDPR (Article 22) prohibits decisions with significant effects on a person — such as being rejected for a job — from being made on a “purely automated” basis, without meaningful human intervention and without the candidate being able to understand the logic and challenge the outcome. The EU AI Act goes further: it classifies recruitment AI as “high-risk”, requiring effective human oversight, bias auditing, and full transparency. And systematically scoring behavioral traits at scale edges dangerously close to what the Regulation calls “social scoring” — a practice that is expressly restricted. Has Bending Spoons published any compliance framework for this? We don’t know. The company has not explained it publicly. And that, in itself, is already a problem.
My second question was, Is this ethical?
This is an even deeper question. A person who has taken the time preparing his/her application, who’s probably written a cover letter with genuine enthusiasm, who might be considering leaving another job to take a chance at Bending Spoons — is rejected by a mathematical model that has never looked him/her in the eye, knows nothing of his/her personal story, and reduces him/her to a set of numerical variables.
To be clear: the problem is not using AI in hiring. The problem is removing humans from the equation entirely. IBM — one of the companies that has gone furthest in integrating AI into its own HR function — offers an instructive guardrail here. Nickle LaMoreaux, IBM’s Chief Human Resources Officer, has been explicit: «I don’t believe that AI can replace humans. We use AI to supplement and aid in decision-making — but not replace it.» IBM’s own framework insists that AI in recruitment should not make hiring decisions but rather help surface candidates who might otherwise go unnoticed and enable recruiters to assess skills at scale — while the actual decision always remains with a human being. IBM further recommends that organizations define, before deploying any AI in HR, not only what they are willing to do with it, but also — critically — what they are unwilling to do, anchoring the entire process in clear ethical principles. This is the right balance: embrace technology to remove bias, increase efficiency, and widen the talent pool, but never surrender the final judgement — the act of truly believing in a candidate — to a model. That last step must remain irreducibly human.
And another question, Is it even good HR?
Here is where the Bending Spoons case gets truly interesting. What makes organizations genuinely resilient and innovative is diversity — of thought, background, personality, and experience. An algorithm trained on the traits of existing high performers will, by design, keep replicating those same traits in every new hire. The result is not a team of exceptional individuals. It is a monoculture — and monocultures are fragile. The greatest breakthroughs rarely come from the candidate who scored highest on every metric. They come from the outlier. The misfit. The one who would have been filtered out in round one.
And then there is the deeper assumption buried in all of this. Even accepting that this process has so far produced a successful company, that does not prove the process caused that success — nor that a completely different group of people, selected by entirely different criteria, could not have achieved the same or even better results. The algorithm encodes one version of «what works.» But there are many versions. Confusing correlation with causation in HR is an old trap. Building an entire hiring philosophy around it — and presenting it as a unique road to excellence — is an extraordinarily confident bet for a company that is, in the end, still quite young. The long-term verdict is nowhere near in.
Summing up, can an algorithm — however sophisticated — capture a person’s true potential? Their capacity to learn, to adapt, to lead when things get hard or their personal circumstances change? Their motivation? Their resilience? Their full, personal story?
Hiring people is, precisely, one of the most human acts that exists inside an organization. Because choosing someone also means believing in someone. And here comes the sharpest irony of this entire model: Bending Spoons system seems to be designed to find the humans who behave most like machines –in a predictive and deterministic way. So why bother looking for humans at all? A robot could eventually do it better. Faster. Cheaper. And without filing a complaint.
Maybe that is Bending Spoons’ next logical step. Or maybe there is something in human beings — their unpredictability, their creativity, their stubborn refusal to be reduced to a score — that no regression model can ever anticipate.
Is that the future we want for our children? To be selected to work in this or that organization by an algorithm? I choose to hold on to the hope that most of us don’t.
Dejar una contestacion