AI-Induced Automation Bias Reduces Mammogram Diagnostic Accuracy

The article explicitly involves an AI system used in medical diagnosis (evaluating mammograms) whose incorrect outputs influenced radiologists' decisions, leading to decreased accuracy in cancer detection. This constitutes indirect harm to health (a), as misdiagnosis can result in delayed or inappropriate treatment. The event stems from the AI system's use and its malfunction (providing wrong categories). Therefore, this qualifies as an AI Incident because the AI system's involvement has directly led to harm (or at least a significant risk of harm) in healthcare decision-making.

The article explicitly involves an AI system used in medical diagnosis (mammogram evaluation) and shows that its incorrect outputs negatively influenced radiologists' decisions, reducing diagnostic accuracy. This constitutes harm to health (a), as misdiagnosis or delayed diagnosis of breast cancer can cause injury or harm to patients. The AI system's use directly led to this harm through automation bias, fulfilling the criteria for an AI Incident.

The article explicitly involves an AI system used as a decision support tool in mammography. The study demonstrates that incorrect AI suggestions directly reduce radiologist accuracy, which constitutes harm to health (a). This harm arises from the use of the AI system and its influence on human decision-making, fulfilling the criteria for an AI Incident. The harm is realized (not just potential), as the AI system's incorrect outputs have already impaired diagnostic performance in the study context.

The article explicitly involves an AI system used as a decision support tool in mammography reading. The study found that when the AI system suggested incorrect BI-RADS categories, radiologists' accuracy significantly decreased, including among experienced professionals. This indicates that the AI system's outputs directly influenced human decision-making, leading to diagnostic errors that can harm patient health. The harm is realized (not just potential), as the AI system's incorrect advice impaired performance, which fits the definition of an AI Incident due to injury or harm to health. The article also discusses the need for safeguards to mitigate this harm, reinforcing the direct link between AI use and adverse outcomes.

The AI system involved is an AI-based decision support tool for mammogram reading, which is explicitly mentioned. The study demonstrates that the AI's incorrect suggestions led to radiologists making more errors in BI-RADS scoring, which is critical for cancer diagnosis and subsequent patient care. This constitutes indirect harm to patients' health due to reliance on faulty AI outputs. Therefore, this event meets the criteria for an AI Incident as the AI system's use has directly led to harm to health (a).

The article explicitly involves an AI system used as a decision support tool in mammography reading. The study demonstrates that when the AI system provides incorrect suggestions, radiologists' performance significantly worsens, even among experienced professionals. This impairment in diagnostic accuracy constitutes harm to health, fulfilling the criteria for an AI Incident. The harm is realized (not just potential), as the AI system's incorrect outputs directly influenced radiologists' decisions, which could lead to patient harm. Hence, this qualifies as an AI Incident rather than a hazard or complementary information.

The article explicitly involves an AI system used as a decision support tool in mammography. The study demonstrates that incorrect AI recommendations caused radiologists, including experienced ones, to make more errors in diagnosis. This constitutes indirect harm to health (harm to persons) because incorrect BI-RADS assessments can lead to misdiagnosis or delayed treatment. Therefore, the event meets the criteria for an AI Incident, as the AI system's use has directly led to harm through impaired diagnostic performance.

The event involves the use of an AI system (AI-based clinical decision support for mammography readings) whose outputs influence radiologists' decisions. The study shows that incorrect AI suggestions lead to significantly worse diagnostic accuracy, which constitutes harm to the health of patients (harm category a). This harm is directly linked to the use of the AI system and its influence on human decision-making, fulfilling the criteria for an AI Incident. The article does not merely discuss potential risks or general information but documents actual observed negative impacts on diagnostic accuracy due to AI use.

The event involves the use of an AI system in medical diagnosis (mammogram reading) and demonstrates that reliance on AI outputs can lead to incorrect clinical assessments, which could directly or indirectly cause harm to patients' health. The study highlights a malfunction in the use of the AI system (automation bias) that affects diagnostic accuracy, a clear link to potential harm to health. Therefore, this qualifies as an AI Incident because the AI system's use has directly led to diagnostic errors that could harm patients.

The event involves the use of an AI system in medical diagnosis (mammogram reading) where the AI's incorrect suggestions caused radiologists to make erroneous assessments. This constitutes indirect harm to health (a) because misdiagnosis can lead to inappropriate treatment or delayed care. The study highlights a malfunction in the use of the AI system (automation bias) affecting decision-making. Therefore, this qualifies as an AI Incident due to realized harm linked to AI system use.

The event involves the use of an AI system in a medical diagnostic context where the AI's incorrect outputs directly influence radiologists' decisions, leading to potential harm to patients through misdiagnosis. The AI system's involvement in the diagnostic process and its erroneous suggestions causing incorrect BI-RADS assessments meet the criteria for an AI Incident, as it has directly led to harm to health (a).

The event involves an AI system used in medical diagnosis (mammogram reading) where the AI's incorrect suggestions caused radiologists to make wrong assessments. This constitutes indirect harm to health because misdiagnosis can lead to inappropriate treatment or delayed care. The study highlights realized harm potential from AI use in healthcare, fitting the definition of an AI Incident due to injury or harm to health caused indirectly by AI system use.

The AI system involved is explicitly mentioned as providing BI-RADS category suggestions for mammograms. The study demonstrates that radiologists often follow incorrect AI suggestions, resulting in a significant decrease in correct mammogram ratings. This constitutes indirect harm to health due to reliance on AI outputs leading to diagnostic errors. Therefore, this event qualifies as an AI Incident because the AI system's use has directly influenced harmful outcomes in medical diagnosis.

The event involves the use of an AI system in medical diagnosis (mammogram reading) where the AI's incorrect suggestions caused radiologists to make erroneous assessments. This constitutes indirect harm to health (a) because misdiagnosis can lead to inappropriate treatment or delayed care. The study highlights a malfunction in the use of the AI system (automation bias) affecting decision-making. Therefore, this qualifies as an AI Incident due to realized harm linked to AI system use.

The article discusses a research study demonstrating that radiologists tend to rely on AI suggestions, even when incorrect, which can reduce diagnostic accuracy. While this indicates a risk of harm due to automation bias, the article does not report a specific incident where harm occurred or was directly caused by the AI system. Instead, it highlights a potential issue that needs safeguards to prevent harm. Therefore, this is best classified as Complementary Information, as it provides important context and understanding about AI's impact in medical diagnostics without describing a realized harm or incident.