A fascinating new study explores whether AI models used in healthcare exhibit biases. Researchers created almost a million fake resumes for medical professionals, tweaking factors like gender and race while keeping qualifications constant. These resumes were then fed to three leading large language models—GPT-4, Claude-3, and Mistral-Large—and asked to rank them for different medical specialties. Surprisingly, all three AIs displayed significant biases. Some favored men for surgery and orthopedics, while others preferred women for family medicine and pediatrics. Racial biases also emerged. While some models leaned towards Asian applicants, GPT-4 showed a preference for Black and Hispanic candidates. Interestingly, the AI's choices often overrepresented minority groups compared to their actual presence in those medical fields. This suggests that these models, trained on vast datasets of text and code, may have absorbed and amplified existing societal biases. While AI holds incredible potential for healthcare, this study highlights the critical need to address these biases before integrating such models into real-world decision-making. Ensuring fairness and avoiding the perpetuation of inequality are crucial for building a just and equitable healthcare system.