Why 78% of 70,000 Candidates Chose an AI Interview

Most people assume that AI in hiring is about replacing human judgment. But a massive new study suggests something else entirely: AI’s greatest value isn't replacing human judgment—it’s replacing human inconsistency.

In a rare, large-scale randomized field experiment (working paper; last revised Jan 26, 2026), researchers from Chicago Booth and Erasmus University studied over 70,000 job applicants. The results offer a blueprint for the future of recruitment.

When given a direct choice between a human recruiter and an AI voice agent—a natural-language system that conducts a phone conversation much like a human—78% of candidates chose the AI.

Here is what we can learn from the data.

1. The Performance Delta

The study focused on high-volume, entry-level roles (customer service and call-center style work) in the Philippines. By randomly assigning candidates to human recruiters or AI voice agents, the researchers found that AI-led screening significantly improved the "quality of hire":

• 12% Lift in Offers: Candidates interviewed by AI were roughly 12% more likely to receive job offers (with offer rates rising from approximately 8.70% to 9.73%).

• The Retention Signal: AI-interviewed candidates were 18% more likely to start the job and 18% more likely to remain employed after 30 days. This durability signal persisted for up to four months.

• No Productivity Drop: Despite the automated screening, hired workers showed no decline in core performance metrics like CSAT (customer satisfaction) or average handle time.

2. The Mechanism: “Controlled Variance”

Why did a machine outperform experienced recruiters at the screening stage? The authors point to a concept called “Controlled Variance.”

Human recruiters—even those following a script—suffer from "noise." They get tired, they have "gut feels," and they vary their questions or tone based on the candidate. This creates a "Recruiter Lottery" where a candidate's success depends on who they talk to.

The AI, however, provides a highly structured and consistent framework while remaining responsive to individual answers. By standardizing the information collection, the AI provided human decision-makers with richer, more comparable data to evaluate.

3. Reducing the "Friction" of Bias

The study also touched on a critical human element: reported discrimination. Candidates in the AI-led group reported significantly lower levels of perceived gender-based discrimination during their interview experience—falling from 5.98% to 3.30%. Candidates chose AI because they felt they would get a "fairer shake" without the subtle micro-judgments that can occur in a human-to-human call.

4. Important Nuance: It’s Not Magic

While these results are compelling, implementation quality matters. The researchers noted that ~7% of AI-led interviews faced technical difficulties. Furthermore, this study focused on high-volume roles where standardized criteria are easier to define; results may differ for executive or highly creative roles.

Crucially, humans still made the final hiring decisions. The AI didn't hire the person; it simply provided a more consistent "first-mile" screen.

The AceUnlock Perspective: Protecting Human Judgment

At AceUnlock, we believe this study validates the next era of recruitment. The goal isn't to remove the human; it's to protect the human.

When recruiters are forced to run 25 repetitive first-round screens a week, they lose the energy and nuance required for high-level evaluation. By using AI for consistency in the "first-mile" screen, we empower recruiters to focus their judgment where it actually matters: the final interview and the offer close.

The takeaway for hiring teams: Stop using your most expensive asset—human judgment—on repetitive data collection. Use AI to surface the signal, and use humans to make the final call.

Dive into the research: Voice AI in Firms: A Natural Field Experiment on Automated Job Interviews (Jabarian & Henkel, 2026)

Or read the full paper: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5395709

Hiring Leaders: Are your recruiters spending their weeks on "The Recruiter Lottery" or on high-value judgment?