The 84% Claude Bias: Why AI Recruiting Tools in Switzerland Favor Their Own Creations

Study reveals AI hiring chatbots exhibit systemic bias. Swiss firms lag in strategic HR digitization despite 90% admin automation, while micro-credentials gain traction.

Swiss companies have pushed over 90 percent of HR administration into the digital age — yet when it comes to strategic tasks like talent development and hiring, manual processes remain the norm. A fresh wave of research now reveals a deeper problem: the artificial intelligence used in recruiting may be subtly rigging the game.

A study by i10x.ai tested how AI chatbots evaluate job applications generated by other AI systems. The results were stark. Claude, the model from Anthropic, recommended its own resume variants for hiring 84 percent of the time. Profiles created using GPT-based tools got a hiring nod only 42 percent of the time. Gemini-generated applications achieved the highest acceptance rate at 94.5 percent. The researchers urge companies to conduct regular bias audits and deploy model panels to prevent one-sided evaluations.

The findings echo broader concerns from the "Digital HR 2026" report by Avenir. While administrative workflows in Swiss firms are nearly fully automated, the picture flips for employee development, performance management and skills tracking — 85 percent of surveyed companies still run these processes manually or partially digital. Recruiting sits in the middle, with a digitalization level of about 65 percent.

Actual AI adoption lags far behind potential. Only one in five companies uses AI in recruiting; just over 10 percent deploy it in administration. The biggest barriers are legal uncertainty (cited by 56 percent of respondents) and missing governance structures (53 percent). Technical implementation troubles rank as a problem for a mere 16 percent.

Human bias also creeps in through the backdoor. The Avenir study found that 47 percent of those surveyed see a significant risk of distortion arising from the preconceptions people embed into AI prompts.

Meanwhile, a quiet shift is underway in how employers evaluate candidates. According to a recent survey, 81 percent of employers already consider micro-credentials — short-form digital certificates — during hiring. Yet 97 percent of HR professionals admitted they do not know the specific concepts involved. After being briefed, 62 percent expressed interest. Two-thirds said integration into existing systems like SAP or Workday is a prerequisite.

Jiaona Zhang of Laurel captured the trend in early June: the focus is moving away from academic majors toward what candidates actually build. What matters increasingly, she said, is how applicants use AI to replace manual work.

The efficiency gains are real. At the company Sana, an AI agent reviewed 1,500 contracts — slashing the time per document from six hours to 15 minutes. Product managers saved up to ten hours per work cycle through automated analysis. But these productivity leaps are reshaping the labor market in jarring ways. In the first five months of 2026, the tech sector cut more than 115,000 jobs. At the same time, demand for "forward-deployed engineers" surged 729 percent between April 2025 and April 2026.

SAP CEO Christian Klein argues that AI needs the operational context of a company — its processes, data and rules. The ultimate aim, he says, is the autonomous enterprise, where AI coordinates execution and humans set strategic direction. In areas like consulting and sales, Klein insists, human intuition — "taste" or strategic thinking — remains indispensable.