AI as expanded artificial intelligence is man’s greatest ally in the modern world. It has promising help, advancements to mankind making tasks easier, providing knowledge in greater depth, and bending as per the rules regulated by humans.

In the medical field, AI enhances diagnostics, treatment planning, and patient care through advanced data analysis and machine learning algorithms. It helps in interpreting medical images like X-rays, MRIs, and CT scans with high accuracy, often detecting conditions that may be missed by human eyes. AI also assists in predicting patient outcomes, personalizing treatment plans based on individual health data, and streamlining administrative tasks such as patient scheduling and electronic health record management. Additionally, AI-driven tools support drug discovery and development by analyzing vast datasets to identify potential therapeutic targets and predict drug efficacy. Until May 2024, the FDA approved 882 AI-enabled medical devices among them 671 is used for diagnosis in radiology.

Artificial intelligence (AI) has become a powerful tool in modern healthcare, particularly in the analysis of medical images. However, a recent study conducted by researchers at the Massachusetts Institute of Technology (MIT) reveals a troubling aspect of this technology: AI models that analyze medical images can produce inconclusive or biased results for women and people of color. This finding underscores the urgent need for greater scrutiny and improvement in the development and implementation of AI systems in healthcare.

The study conducted by MIT researchers focused on evaluating the performance of AI models used in medical imaging across diverse patient demographics. The researchers discovered that these models, which are often trained on datasets that lack sufficient representation of women and people of color, tend to be less accurate and reliable for these groups. This is primarily because many of these models are trained on datasets that are predominantly composed of male patients. Consequently, the AI systems struggle to accurately interpret medical images from female patients, leading to higher rates of misdiagnosis or uncertain outcomes. The study also revealed a significant bias against people of color in AI-generated medical image analyses. Similar to the issue with gender, the underrepresentation of people of color in training datasets results in AI models that are less effective at diagnosing and interpreting medical conditions for these populations. This bias can lead to poorer health outcomes for people of color and women, as their medical needs may not be accurately identified or treated by AI-driven diagnostic tools.

Certain key factors exposed the AI models’ bias if the detection was repeated on the same subject and the data detection was correctly predicted, this becomes dangerous because in a lot of cases, data recorded by other hospitals is used for the detection of disease and if the AI generates inconclusiveness, then there will be grave misconduct.

By addressing these biases through diversified training data, rigorous testing, continuous monitoring, and collaborative efforts, the healthcare industry can harness the full potential of AI to deliver equitable and effective care for all patients. As we move forward, it is imperative to prioritize inclusivity and fairness in the development and deployment of AI technologies in healthcare. The study was funded by the Google Research Scholar Award, the Robert Wood Johnson Foundation Harold Amos Medical Faculty Development Program, RSNA Health Disparities, the Lacuna Fund, the Gordon and Betty Moore Foundation, the National Institute of Biomedical Imaging and Bioengineering, and the National Heart, Lung, and Blood Institute, helped in introducing a caution among hospitals.