AI Bias News

Latest Articles

• Clinical AI Diagnostic Accuracy Plummets in New AgentClinic Benchmark Mimicking Real-World Patient Uncertainty

  Published: May 1, 2026

  Section: Health Technology

  A new multi-modal benchmark called AgentClinic reveals that AI models which excel at static medical exams struggle when faced with interactive, sequential patient simulations. Whil...

• Virginia Tech research finds AI models reinforce autism stereotypes when providing social advice

  Published: Apr 20, 2026

  Section: Mental Health

  A study presented at the 2026 CHI conference reveals that large language models (LLMs) frequently give restrictive and stereotypical advice to users who disclose an autism diagnosi...

• Virginia Tech researchers find AI models discourage social interaction for autistic users based on ingrained stereotypes

  Published: Apr 17, 2026

  Section: Science & Tech

  A study led by Virginia Tech doctoral student Caleb Wohn reveals that major artificial intelligence models frequently provide social advice rooted in autism stereotypes rather than...

• Nvidia Faces Backlash as New AI Rendering Tool Automatically Enhances Female Character Aesthetics

  Published: Apr 4, 2026

  Section: Science & Tech

  The launch of Nvidia’s DLSS-5 technology has ignited a controversy over AI-driven "beauty filters" that alter video game character models without developer consent. While the compa...

• Cornell Tech Study Reveals AI Writing Tools Exert Covert Influence on Human Opinions Despite Explicit Warnings of Algorithmic Bias

  Published: Mar 13, 2026

  Section: Research

  Researchers at Cornell Tech have discovered that AI-powered autocomplete suggestions possess the power to shift a writer's personal attitudes on sensitive societal issues. In large...

• Brown University research identifies fifteen distinct ethical risks in the use of AI chatbots for mental health counseling

  Published: Mar 4, 2026

  Section: Health

  A study involving computer scientists and clinical psychologists has found that large language models frequently violate professional ethical standards when providing therapy-style...

• Inside the Black Box: MIT Researchers Develop Method to Expose and ‘Steer’ Hidden Personas in Large Language Models

  Published: Feb 26, 2026

  Section: Science & Tech

  A research team from MIT and UC San Diego has developed a groundbreaking method to identify and manipulate abstract concepts—such as moods, biases, and personalities—hidden within...