Smart Criteria Nudge

Smart Criteria Nudge changes the order of information displayed in the AI-powered recruiting dashboard. Instead of recruiters seeing an AI match score right away, four additional steps are inserted.
First, they must actively confirm that they are conducting the evaluation based on predefined criteria (transparency check). They then manually evaluate five weighted competencies (Professional Qualifications 30%, Problem-Solving Skills 20%, Teamwork 20%, Communication 15%, Initiative 15%). They are then asked to weigh the arguments for and against the applicants’ suitability (reflection prompt). Only then is the AI Match Score revealed, serving as a secondary decision-making aid rather than a mental anchor.

Accept the cookie setting to see Panopto content.

Unlock Panopto

What does the topic mean?

51% of companies now use AI tools in recruiting, and about half use large language models (LLMs) for automated screening of applications.
The problem is that LLMs are not neutral entities. They learn from human datasets and internalize those biases. If AI scores are displayed at the very beginning of the screening process, they serve as a mental reference point and distort subsequent human judgment. Recruiters then tend to accept the AI assessment uncritically, rather than evaluating it on their own.

Goal of the nudge

By changing the order of the information, the goal is to interrupt intuitive “quick judgment” (System 1) and encourage slower, criteria-based thinking (System 2). The final decision should consciously remain with the person.

    Needs analysis

    • Widespread Use of AI in Recruiting
      51% of companies use AI tools, 44% of which are specifically for automated screening. Integration is accelerating rapidly.
    • Lack of Safeguards
      Current recruiting platforms typically display AI scores immediately. There is no mechanism to separate human evaluation from AI influence.
    • Trend Toward Automation
      Recruiters are increasingly relying on AI recommendations without critically evaluating them, a phenomenon known as "cognitive surrender."

    Cause analysis

    • Anchoring Bias

      The AI score displayed early on becomes the dominant mental anchor. Recruiters do not
      sufficiently adjust their judgment away from it.

    • Confirmation Bias

      Once anchored, recruiters interpret information in a way that confirms the AI assessment. In a study involving 775 managers, Carter and Liu (2025) demonstrated that the AI score anchors the judgment so strongly that subsequent information is processed primarily in a way that confirms it.

    • Framing Effects

      Even minimal changes in the wording of a prompt can shift an LLM's evaluation by several
      categories.

Target Group

The Nudge is aimed at recruiters and HR professionals in organizations that use AI-powered
recruiting tools or are planning to implement them.

Added value of the nudge

  • Better Decision-Making

    Cognitive biases are reduced by structuring the evaluation process and delaying the integration of
    AI.

  • Transparency

    The transparency check and the reflection question promote a more conscious approach to AIbased recommendations. They enhance the transparency of the decision-making process and
    encourage a critical examination of AI assessments.

  • Fairer Hiring Practices

    By reducing the systematic transfer of bias from AI to human recruiters.

CONTACT US

Evelin Schende

evelin.schendel@stud.hshl.de

Emily Schmidt

emily.schmidt@stud.hshl.de

Jolin Niemann

jolin.niemann@stud.hshl.de