Business A.M
  • Finance & Investment
  • Company & Business
    • Insurance & Pension Business
  • Technology
  • Commodities
  • Comments
    • Analysts Insights
  • E-Copy
Top Posts
Sahara Group boosts environmental sustainability, economic empowerment with...
The burden of a virus
States generate N986.2 billion IGR in 9-month, rise...
Nigeria Daily Street Market Exchange Rate (December 20th...
Sahara Group, Amigo LNG collaborate to boost global...
McKinsey shows how insurers can rethink SME segment
Pension investments in FG securities up 20.6% to...
How We Can Harness AI to Fulfill Our...
Experts urge banks to make SMS alerts optional as...
Third-Party cyber risks fuel rising insurance claims,financial losses
AEC Chair Ayuk urges stronger role for local...
Nigeria Stock Exchange surges past Nairobi, J’burg in...
Policy focus: Prosperity creation or poverty alleviation?
BUA Cement to expand Sokoto plant with $240m...
Sunday, January 25, 2026
Business A.M
SUBSCRIBE
  • Finance & Investment
  • Company & Business
    • Insurance & Pension Business
  • Technology
  • Commodities
  • Comments
    • Analysts Insights
  • E-Copy
Top Posts
Sahara Group boosts environmental sustainability, economic empowerment with...
The burden of a virus
States generate N986.2 billion IGR in 9-month, rise...
Nigeria Daily Street Market Exchange Rate (December 20th...
Sahara Group, Amigo LNG collaborate to boost global...
McKinsey shows how insurers can rethink SME segment
Pension investments in FG securities up 20.6% to...
How We Can Harness AI to Fulfill Our...
Experts urge banks to make SMS alerts optional as...
Third-Party cyber risks fuel rising insurance claims,financial losses
AEC Chair Ayuk urges stronger role for local...
Nigeria Stock Exchange surges past Nairobi, J’burg in...
Policy focus: Prosperity creation or poverty alleviation?
BUA Cement to expand Sokoto plant with $240m...
Sunday, January 25, 2026
Business A.M
  • Finance & Investment
  • Company & Business
    • Insurance & Pension Business
  • Technology
  • Commodities
  • Comments
    • Analysts Insights
  • E-Copy
Copyright 2021 - All Right Reserved
Home » How to Detect Bias in Large Language Models
Knowledge@Wharton

How to Detect Bias in Large Language Models

by KNOWLEDGE WHARTON December 10, 2025
by KNOWLEDGE WHARTON December 10, 2025 0 comments 1.8K views 3 minutes read Share
TwitterPinterestTumblrVKOdnoklassnikiRedditStumbleuponLINEPocketSkypeViberThreadsBlueskyEmail
1.8K

Research from Wharton’s Sonny Tambe finds that LLMs can make biased hiring decisions that traditional auditing methods might not be able to catch.

As AI becomes a fixture in hiring, evaluation, and policy decisions, a new study funded by the Wharton AI & Analytics Initiative offers a rigorous look at a critical question: Do race and gender shape how large language models (LLMs) evaluate people? If so, how can we tell?
The answer, according to Prasanna (Sonny) Tambe, faculty co-director of Wharton Human AI Research, and others, is complex, and the implications matter for every organization deploying LLMs at scale. Here are the key takeaways you need to know from Tambe’s latest research on LLM bias and auditability.

Bias isn’t just a human problem — it shows up in the code.
Despite their veneer of neutrality, LLMs trained on vast swaths of online data can absorb and replicate human biases. This study shows that when prompted with the application materials of job candidates, LLMs systematically produced different evaluations depending on whether a person was described as Black, Hispanic, Asian, or White, and whether they were male or female, even when everything else was kept the same.
The direction of these biases is not always predictable. For example, the LLMs tested in the study rated women and people of color more favorably than White men, a reversal of traditional discrimination patterns. But researchers caution against assuming this is a “fairness fix.” It may signal overcorrection in post-training intended to correct biases, which can generate its own kind of undesirable effects.

Auditing LLMs requires new methods, not just old metrics.
Traditional evaluation methods weren’t enough to diagnose bias in this study. Adverse impact ratio, a widely used auditing metric, showed some disparities, but the results were too imprecise to draw strong conclusions. That’s why Tambe and his colleagues pioneered a new approach: LLM-based correspondence experiments.
Inspired by methods used to detect discrimination in human hiring, these experiments carefully manipulated résumés and interview transcripts. By changing only names and pronouns to signal race and gender, the team could measure how models respond to applicants with identical qualifications across demographic lines.

The disparities are subtle, but persistent and meaningful.
Using this method across 11 top LLMs from OpenAI, Anthropic, and Mistral, researchers found that women and racial minorities received slightly higher ratings than their White male counterparts. The differences were modest — often just a few percentage points.

These results held even when researchers:
Changed the district context from diverse to predominantly White
Altered the evaluation prompts
Removed interview transcripts, relying on résumés alone.
That robustness suggests the disparities are embedded in how the models were trained or aligned, not just a response to specific prompt wording or context.
Audits must match the use case, and context matters.
The research stresses that LLM bias can’t be fully understood outside its application. A model may behave differently depending on task, prompt, or population. For example, tools used for hiring may need different audits than those used in customer service or credit risk evaluation.
Auditing LLMs isn’t one-size-fits-all. Policymakers and organizations need context-specific audits to understand how these models actually perform in the real world.

LLM audits are essential infrastructure for ethical AI.
This study isn’t just meant to be a warning — it also offers a roadmap. Tambe and his colleagues provide companies, researchers, and regulators with a powerful tool to hold language models accountable in evaluation contexts. In doing so, they help ensure AI deployment aligns with legal standards and social expectations.
As Tambe explains, “what makes this problem urgent is how widespread LLM use already is becoming in organizational workflows. And yet, we don’t yet have robust standards for understanding how these models perform with respect to fairness.”

Bottom line: Don’t deploy LLMs blindly. Audit them.
Organizations are rushing to integrate LLMs into decision-making pipelines. This research is a timely reminder: Even the smartest models aren’t immune to bias. But with the right tools, we can ensure their outputs are just.

You Might Also Like
  • How Diverse Boards Can Help Close the Wealth Gap
  • What Machine Learning Reveals About Forming a Healthy Habit
  • How Vaccine Mandates Are Helping Companies
  • What Is Common Sense?
Share 0 TwitterPinterestTumblrVKOdnoklassnikiRedditStumbleuponLINEPocketSkypeViberThreadsBlueskyEmail
previous post
How Can Leaders Adapt to AI?
next post
Glo’s connectivity promise propels Nigeria into faster future

related posts

Onboarding Best Practices: Maximum Value From New Hires

January 9, 2026

How the Global South Can Chart a New Path for AI

January 9, 2026

Adaptive Experimentation in the Age of AI

January 2, 2026

Can Your Face Predict Your Salary? Using AI Personality Assessments in Hiring

January 2, 2026

Look Them in the Eye: Strengthen Teams and Trust

December 27, 2025

Why Is Your Credit Card Rate So High?

December 27, 2025

Recent Posts

  • Anambra oil bearing communities shut down 12 SNEEPCO wells over 10 yrs neglect

    January 23, 2026
  • Taraba, with N2.04trn economy, struggles for recognition, development

    January 23, 2026
  • WHO sugar tax drive meets global resistance

    January 23, 2026
  • Market pullback erases N557bn as investors lock in gains on Nigerian bourse

    January 23, 2026
  • BUA Cement to expand Sokoto plant with $240m investment

    January 22, 2026
  • Paymentology taps Kesheni Moodley to drive Africa expansion

    January 22, 2026
  • CPPE warns higher sugar taxes could hurt Nigeria’s industrial recovery

    January 22, 2026
  • NECA seeks deeper collaboration to improve enforcement of employee compensation law

    January 22, 2026
  • ICBA challenges WHO push for higher sugar drink taxes

    January 22, 2026
  • Microsoft tops phishing hit list as cybercriminals hunt login credentials

    January 22, 2026

Business a.m. is Nigeria’s leading business and financial newspaper, delivering in-depth news, analysis, and insights on economics, markets, finance, technology, agriculture, and policy. With a commitment to quality journalism and data-driven reporting,

Facebook Twitter Youtube Linkedin Envelope Rss

Useful Links

    • Work With Us
    • Contact Us
    • Collaboration
    • Data Collection
    • Workplace
    • Adverstising
    • Privacy Policy
    • International Collab
    • Feedback
    • Terms of Use
    • About Our Ads
    • Help & Support
    • Entertainment
    • News Covering
    • Technology
    • Trending Now

Edtior's Picks

Anambra oil bearing communities shut down 12 SNEEPCO wells over 10 yrs neglect
Taraba, with N2.04trn economy, struggles for recognition, development
WHO sugar tax drive meets global resistance

Latest Articles

Anambra oil bearing communities shut down 12 SNEEPCO wells over 10 yrs neglect
Taraba, with N2.04trn economy, struggles for recognition, development
WHO sugar tax drive meets global resistance
Market pullback erases N557bn as investors lock in gains on Nigerian bourse

Businessamlive 2025

  • Finance & Investment
  • Company & Business
    • Insurance & Pension Business
  • Technology
  • Commodities
  • Comments
    • Analysts Insights
  • E-Copy
Business A.M
  • Home
  • HOT Categories
    • Politics
    • Covid-19 HOT
    • European Union
    • City & Business
    • Sport Daily
    • Personal Finance
    • Example Item
    • Example Item
    • Example Item
    • Example Item
  • Entertainment
    • TV & Radio
    • Celebrity News
    • Gaming
    • Hot Music Hot
    • Films Review Hot
    • Express Wins
    • Example Item
    • Example Item
    • Example Item
    • Example Item

Popular This Week

New Metaverse technology allows you to kiss and feel...
by Chris
How tech companies are trying to woo employees returning...
by Chris
Gold rallies past $3,390 as market reacts to sliding...
by Onome Amuge
Nigeria Daily Street Market Exchange Rate (December 20th 2019)
by Chris

Editor's Pick

Cryptocurrency cyber crime has cost victims millions this year
by Chris
Airtel sees 11.5% revenue growth to $5.26bn in ‘23
by Chris
The West Is Not Dying, but It Is Working on...
by Chris
CBN forecasts rise in secured credit lending to households, small...
by Chris