Michael A. Bruno
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Michael A. Bruno has a total of 135 co-authors. Here are the top co-authors by number of shared publications:
- Grayson L. Baird
Brown University
6 shared publications Elizabeth A. Krupinski
Department of Radiology and Imaging Sciences , School of Medicine, Emory University , Atlanta , GA , USA
5 shared publications- Brian Sheppard
The Brown Radiology, Psychology, and Law Lab
4 shared publications - Michael H. Bernstein
Brown University
4 shared publications Hani H. Abujudeh
Detroit Medical Center/Wayne State University and Envision Corporation, Detroit, Michigan
3 shared publicationsJonelle Petscavage-Thomas
Department of Radiology, Penn State Milton S. Hershey Medical Center and the Penn State College of Medicine, Hershey, PA, United States of America
3 shared publications
Latest Publications
- Open Access
- 13 August 2026
- Published by Walter de Gruyter GmbH in Diagnosis
Enhancing diagnostic safety: addressing knowledge gaps for using human factors tools in the safe and effective use of AI – a proposed research agenda
Introduction: Identify knowledge gaps in applying artificial intelligence in clinical settings, using medical imaging as a primary use case to enhance diagnostic efficacy, efficiency, and patient and provider safety. Methods: We convened a two-day workshop with 18 interdisciplinary experts from three countries. Experts represented quality and patient safety, human factors and systems engineering, radiology and other medical specialties, nursing, medical informatics, cognitive and perceptual psychology, psychometrics and statistics, and machine learning, drawn from academia, industry, health systems, and government. Results: We identified by consensus six major knowledge-gap domains, with specific research questions for each domain: development, validation, integration and sustainability; redesign of existing healthcare systems; human and team augmentation; deployment of adaptive-learning “Foundation Models;” and balancing innovation, standardization, and regulatory oversight. Conclusions: We recommend employing a multidisciplinary collaborative approach in future research to leverage transformational AI capabilities anticipated in the next 5–7 years for each of the identified knowledge-gap domains, including ensuring that clinical AI supports diagnostic decision-making, integrates into clinical workflows, and mitigates risks related to automation bias, overreliance, fragmentation of care, and unintended consequences.
- 13 August 2026
- Published by Walter de Gruyter GmbH in Diagnosis
Open AccessEnhancing diagnostic safety: addressing knowledge gaps for using human factors tools in the safe and effective use of AI – a proposed research agenda
Introduction: Identify knowledge gaps in applying artificial intelligence in clinical settings, using medical imaging as a primary use case to enhance diagnostic efficacy, efficiency, and patient and provider safety. Methods: We convened a two-day workshop with 18 interdisciplinary experts from three countries. Experts represented quality and patient safety, human factors and systems engineering, radiology and other medical specialties, nursing, medical informatics, cognitive and perceptual psychology, psychometrics and statistics, and machine learning, drawn from academia, industry, health systems, and government. Results: We identified by consensus six major knowledge-gap domains, with specific research questions for each domain: development, validation, integration and sustainability; redesign of existing healthcare systems; human and team augmentation; deployment of adaptive-learning “Foundation Models;” and balancing innovation, standardization, and regulatory oversight. Conclusions: We recommend employing a multidisciplinary collaborative approach in future research to leverage transformational AI capabilities anticipated in the next 5–7 years for each of the identified knowledge-gap domains, including ensuring that clinical AI supports diagnostic decision-making, integrates into clinical workflows, and mitigates risks related to automation bias, overreliance, fragmentation of care, and unintended consequences.
- Open Access
- 21 July 2026
- Published by SAGE Publications in Neuroscience Insights
- Vol. 21
Neurocognitive Mechanism of Perceptual Errors in Radiologists: Results of Pilot Studies
Background: The most prevalent type of radiologist error is failing to detect abnormalities on images, the so-called “perceptual error.” The prevalence of this type of false-negative (FN) error remains essentially unchanged since it was first described in 1949. Reducing this error rate would be highly beneficial for patients who often depend on radiological observations for a timely diagnosis. Purpose: To identify a potential neurocognitive mechanism contributing to radiologists’ perceptual errors, in order to inform intervention strategies to reduce such errors in practice. These experiments evaluated the relationship between brain network activation states and radiologists’ perceptual errors on two distinct visual tasks utilizing functional MRI (fMRI) and functional Near Infrared Spectroscopy (fNIRS). Materials and Methods: A prospective, pilot study was carried out on a small number of radiologist participants utilizing two distinct types of visual tasks while monitoring activity in two targeted brain networks, the default mode network (DMN) and the “executive” frontoparietal network (FPN), to test the hypothesis that simultaneous co-activation of these networks produces an error-prone neurocognitive state (EPS). A third experiment using fNIRS alone was an observational study of subjects’ neurocognitive states during their usual practice to determine the baseline prevalence of the EPS in the clinical setting. Results: An approximately threefold increased risk of FN perceptual errors (misses) on the visual tasks was observed in the presence of simultaneous co-activation of elements of the Default Mode Network (DMN) and Frontoparietal Network (FPN), which supports the “error-prone state” (EPS) hypothesis. EPS prevalence declined with participant age. Conclusion: Our results suggest that dynamic interactions between the Default Mode Network (DMN) and Frontoparietal Network (FPN) create a unique error-prone state (EPS), associated with an approximately threefold increased risk of perceptual error in radiologists. These results suggest potential intervention strategies for perceptual error, the largest class of radiologist errors in practice.
- 21 July 2026
- Published by SAGE Publications in Neuroscience Insights
- Vol. 21
Open AccessNeurocognitive Mechanism of Perceptual Errors in Radiologists: Results of Pilot Studies
Background: The most prevalent type of radiologist error is failing to detect abnormalities on images, the so-called “perceptual error.” The prevalence of this type of false-negative (FN) error remains essentially unchanged since it was first described in 1949. Reducing this error rate would be highly beneficial for patients who often depend on radiological observations for a timely diagnosis. Purpose: To identify a potential neurocognitive mechanism contributing to radiologists’ perceptual errors, in order to inform intervention strategies to reduce such errors in practice. These experiments evaluated the relationship between brain network activation states and radiologists’ perceptual errors on two distinct visual tasks utilizing functional MRI (fMRI) and functional Near Infrared Spectroscopy (fNIRS). Materials and Methods: A prospective, pilot study was carried out on a small number of radiologist participants utilizing two distinct types of visual tasks while monitoring activity in two targeted brain networks, the default mode network (DMN) and the “executive” frontoparietal network (FPN), to test the hypothesis that simultaneous co-activation of these networks produces an error-prone neurocognitive state (EPS). A third experiment using fNIRS alone was an observational study of subjects’ neurocognitive states during their usual practice to determine the baseline prevalence of the EPS in the clinical setting. Results: An approximately threefold increased risk of FN perceptual errors (misses) on the visual tasks was observed in the presence of simultaneous co-activation of elements of the Default Mode Network (DMN) and Frontoparietal Network (FPN), which supports the “error-prone state” (EPS) hypothesis. EPS prevalence declined with participant age. Conclusion: Our results suggest that dynamic interactions between the Default Mode Network (DMN) and Frontoparietal Network (FPN) create a unique error-prone state (EPS), associated with an approximately threefold increased risk of perceptual error in radiologists. These results suggest potential intervention strategies for perceptual error, the largest class of radiologist errors in practice.
- Open Access
An Experimental Investigation of the Relationship between AI-Human Workflow Design and Legal Liability for Radiologists: The Erroneous-Change Penalty and Omission Bias
Background: With growing impetus to integrate artificial intelligence (AI) tools into radiology, clinical practices must navigate workflow redesign. This carries implications for medical malpractice liability. Methods: We conducted an online vignette experiment with United States adults who acted as hypothetical jurors in a malpractice case involving a missed intracranial hemorrhage. Participants (n=2,347) were randomized to one of 22 conditions: a no-AI control and 21 conditions involving a hypothetical AI system. These twenty-one conditions varied by whether (1) a single-read or double-read workflow was used, (2) the radiologist’s initial interpretation was documented, (3) the radiologist changed their interpretation after viewing AI output, (4) the AI detected the abnormality, and (5) the AI error rate—False Discovery Rate (FDR) or False Omission Rate (FOR)—was provided to participants only, both participants and radiologist, or neither. The primary outcome was perceived liability, assessed by whether the radiologist met their duty of care. Findings: Perceived liability differed across conditions (p<0.0001). Double-read workflows (p<0.0001), documenting initial interpretations (p=0.0125), and providing participants with AI error rates, including the FDR (p=0.0038) or FOR (p=0.0035), reduced perceived liability. Liability was also lower when AI was incorrect (p<0.0001). Radiologists’ awareness of AI error rates did not significantly impact liability. Notably, we observed an “erroneous change penalty”: the greatest liability occurred when radiologists initially identified an abnormality but later changed their interpretation to normal after seeing that AI identified the case as normal; conversely, perceived liability was lowest with documented, double-read workflows. Interpretation: Double-read workflows with documented initial interpretations and disclosure of AI error rates reduce perceived liability, though changing a correct initial interpretation increases it. Strategic workflow design is critical for successful AI implementation that can mitigate malpractice risk. RESEARCH IN CONTEXT: Evidence before this study: Emerging research has identified several factors that shape how artificial intelligence (AI) systems are integrated into clinical workflows. Beyond technical performance, factors such as disease prevalence, documentation practices, and workflow design have all been shown to play a role in the implementation of AI tools and how they will inevitably affect physician liability. Vignette experiments have separately identified cognitive biases and mitigating measures that can be integrated into workflow design, such as disclosing AI error rates and documenting independent interpretations before reviewing AI output. Added value of this study: This study builds on prior work by examining how multiple aspects of AI implementation influence perceptions of legal liability when hypothetical jurors are asked to adjudicate a medical malpractice lawsuit arising from a false negative interpretation. In particular, we show how combining double-read workflows, documentation practices, and the inclusion of AI error rates can reduce perceived liability. We also show that, when a double-read workflow is used, changing an initially correct interpretation to an incorrect one incurs greater liability for the radiologist than being incorrect in both the initial and final interpretations. These findings underscore the need to address cognitive biases that will undoubtedly arise at the human-AI interface. Implications of all the available evidence: Optimizing the integration of AI tools into radiology requires strategic attention to workflow redesign, as combinations of features can collectively affect perceived liability, likely through well-known cognitive biases. Based on the results of our study, we propose one workflow that can mitigate a radiologist’s risk of legal liability. Moving forward, clinical practices and stakeholders should remain cognizant of these factors as they work toward building sustainable AI-physician systems.
Open AccessAn Experimental Investigation of the Relationship between AI-Human Workflow Design and Legal Liability for Radiologists: The Erroneous-Change Penalty and Omission Bias
Background: With growing impetus to integrate artificial intelligence (AI) tools into radiology, clinical practices must navigate workflow redesign. This carries implications for medical malpractice liability. Methods: We conducted an online vignette experiment with United States adults who acted as hypothetical jurors in a malpractice case involving a missed intracranial hemorrhage. Participants (n=2,347) were randomized to one of 22 conditions: a no-AI control and 21 conditions involving a hypothetical AI system. These twenty-one conditions varied by whether (1) a single-read or double-read workflow was used, (2) the radiologist’s initial interpretation was documented, (3) the radiologist changed their interpretation after viewing AI output, (4) the AI detected the abnormality, and (5) the AI error rate—False Discovery Rate (FDR) or False Omission Rate (FOR)—was provided to participants only, both participants and radiologist, or neither. The primary outcome was perceived liability, assessed by whether the radiologist met their duty of care. Findings: Perceived liability differed across conditions (p<0.0001). Double-read workflows (p<0.0001), documenting initial interpretations (p=0.0125), and providing participants with AI error rates, including the FDR (p=0.0038) or FOR (p=0.0035), reduced perceived liability. Liability was also lower when AI was incorrect (p<0.0001). Radiologists’ awareness of AI error rates did not significantly impact liability. Notably, we observed an “erroneous change penalty”: the greatest liability occurred when radiologists initially identified an abnormality but later changed their interpretation to normal after seeing that AI identified the case as normal; conversely, perceived liability was lowest with documented, double-read workflows. Interpretation: Double-read workflows with documented initial interpretations and disclosure of AI error rates reduce perceived liability, though changing a correct initial interpretation increases it. Strategic workflow design is critical for successful AI implementation that can mitigate malpractice risk. RESEARCH IN CONTEXT: Evidence before this study: Emerging research has identified several factors that shape how artificial intelligence (AI) systems are integrated into clinical workflows. Beyond technical performance, factors such as disease prevalence, documentation practices, and workflow design have all been shown to play a role in the implementation of AI tools and how they will inevitably affect physician liability. Vignette experiments have separately identified cognitive biases and mitigating measures that can be integrated into workflow design, such as disclosing AI error rates and documenting independent interpretations before reviewing AI output. Added value of this study: This study builds on prior work by examining how multiple aspects of AI implementation influence perceptions of legal liability when hypothetical jurors are asked to adjudicate a medical malpractice lawsuit arising from a false negative interpretation. In particular, we show how combining double-read workflows, documentation practices, and the inclusion of AI error rates can reduce perceived liability. We also show that, when a double-read workflow is used, changing an initially correct interpretation to an incorrect one incurs greater liability for the radiologist than being incorrect in both the initial and final interpretations. These findings underscore the need to address cognitive biases that will undoubtedly arise at the human-AI interface. Implications of all the available evidence: Optimizing the integration of AI tools into radiology requires strategic attention to workflow redesign, as combinations of features can collectively affect perceived liability, likely through well-known cognitive biases. Based on the results of our study, we propose one workflow that can mitigate a radiologist’s risk of legal liability. Moving forward, clinical practices and stakeholders should remain cognizant of these factors as they work toward building sustainable AI-physician systems.
- 29 January 2026
- Published by Elsevier in Journal of the American College of Radiology
Potential for Expansion of Existing US Residency Programs in Diagnostic Radiology
- 29 January 2026
- Published by Elsevier in Journal of the American College of Radiology
Potential for Expansion of Existing US Residency Programs in Diagnostic Radiology
- Open Access
- 15 January 2026
- Published by Patient Safety Authority in Patient Safety
- Vol. 8  (2)
Artificial or Intelligent? The Impact of AI on Academic Publishing
- 15 January 2026
- Published by Patient Safety Authority in Patient Safety
- Vol. 8  (2)
Open AccessArtificial or Intelligent? The Impact of AI on Academic Publishing
- 5 August 2025
- Published by Elsevier in Journal of the American College of Radiology
- Vol. 22  (11) , 1417-1422
Artificial Intelligence and Its Impact on Radiology: Summary of the 2024 Intersociety Summer Conference
- 5 August 2025
- Published by Elsevier in Journal of the American College of Radiology
- Vol. 22  (11) , 1417-1422
Artificial Intelligence and Its Impact on Radiology: Summary of the 2024 Intersociety Summer Conference
- 20 July 2025
- Published by Elsevier in Academic Radiology
- Vol. 32, S154-S158
The Impact of Cognitive Bias in Radiology Lawsuits
- 20 July 2025
- Published by Elsevier in Academic Radiology
- Vol. 32, S154-S158
The Impact of Cognitive Bias in Radiology Lawsuits
- Open Access
Radiologist-AI workflow can be modified to reduce the risk of medical malpractice claims
Background Artificial Intelligence (AI) is rapidly changing the legal landscape of radiology. Results from a previous experiment suggested that providing AI error rates can reduce perceived radiologist culpability, as judged by mock jury members (4). The current study advances this work by examining whether the radiologist’s behavior also impacts perceptions of liability. Methods. Participants (n=282) read about a hypothetical malpractice case where a 50-year-old who visited the Emergency Department with acute neurological symptoms received a brain CT scan to determine if bleeding was present. An AI system was used by the radiologist who interpreted imaging. The AI system correctly flagged the case as abnormal. Nonetheless, the radiologist concluded no evidence of bleeding, and the blood-thinner t-PA was administered. Participants were randomly assigned to either a 1.) single-read condition, where the radiologist interpreted the CT once after seeing AI feedback, or 2.) a double-read condition, where the radiologist interpreted the CT twice, first without AI and then with AI feedback. Participants were then told the patient suffered irreversible brain damage due to the missed brain bleed, resulting in the patient (plaintiff) suing the radiologist (defendant). Participants indicated whether the radiologist met their duty of care to the patient (yes/no). Results. Hypothetical jurors were more likely to side with the plaintiff in the single-read condition (106/142, 74.7%) than in the double-read condition (74/140, 52.9%), p=0.0002. Conclusion. This suggests that the penalty for disagreeing with correct AI can be mitigated when images are interpreted twice, or at least if a radiologist gives an interpretation before AI is used.
Open AccessRadiologist-AI workflow can be modified to reduce the risk of medical malpractice claims
Background Artificial Intelligence (AI) is rapidly changing the legal landscape of radiology. Results from a previous experiment suggested that providing AI error rates can reduce perceived radiologist culpability, as judged by mock jury members (4). The current study advances this work by examining whether the radiologist’s behavior also impacts perceptions of liability. Methods. Participants (n=282) read about a hypothetical malpractice case where a 50-year-old who visited the Emergency Department with acute neurological symptoms received a brain CT scan to determine if bleeding was present. An AI system was used by the radiologist who interpreted imaging. The AI system correctly flagged the case as abnormal. Nonetheless, the radiologist concluded no evidence of bleeding, and the blood-thinner t-PA was administered. Participants were randomly assigned to either a 1.) single-read condition, where the radiologist interpreted the CT once after seeing AI feedback, or 2.) a double-read condition, where the radiologist interpreted the CT twice, first without AI and then with AI feedback. Participants were then told the patient suffered irreversible brain damage due to the missed brain bleed, resulting in the patient (plaintiff) suing the radiologist (defendant). Participants indicated whether the radiologist met their duty of care to the patient (yes/no). Results. Hypothetical jurors were more likely to side with the plaintiff in the single-read condition (106/142, 74.7%) than in the double-read condition (74/140, 52.9%), p=0.0002. Conclusion. This suggests that the penalty for disagreeing with correct AI can be mitigated when images are interpreted twice, or at least if a radiologist gives an interpretation before AI is used.
- 22 May 2025
- Published by Massachusetts Medical Society in NEJM AI
- Vol. 2  (6)
Randomized Study of the Impact of AI on Perceived Legal Liability for Radiologists
- 22 May 2025
- Published by Massachusetts Medical Society in NEJM AI
- Vol. 2  (6)
Randomized Study of the Impact of AI on Perceived Legal Liability for Radiologists
- 31 October 2024
- Published by Elsevier in Journal of the American College of Radiology
- Vol. 22  (1) , 129-138
Seven Challenges in Radiology Practice: From Declining Reimbursement to Inadequate Labor Force: Summary of the 2023 ACR Intersociety Meeting
- 31 October 2024
- Published by Elsevier in Journal of the American College of Radiology
- Vol. 22  (1) , 129-138
Seven Challenges in Radiology Practice: From Declining Reimbursement to Inadequate Labor Force: Summary of the 2023 ACR Intersociety Meeting