An official website of the United States government
About Ethics, Bias, and Transparency for People and Machines
Artificial intelligence and machine learning (AI/ML) are a collection of data-driven technologies with the potential to significantly advance scientific discovery in biomedical and behavioral research. Researchers employing these technologies must take steps to minimize the harms that could result from their research, including but not limited to addressing (1) biases in datasets, algorithms, and applications; (2) issues related to identifiability and privacy; (3) impacts on disadvantaged or marginalized groups; (4) health disparities; and (5) unintended, adverse social, individual, and community consequences of research and development. Some of the inherent characteristics of AI/ML, as well as its relative newness in the biomedical and behavioral sciences, have made it difficult for researchers to apply ethical principles in the development and use of AI/ML, particularly for basic research.
To address these issues, National Institutes of Health (NIH) Office of Data Science Strategy (ODSS) announced “Administrative Supplements for Advancing the Ethical Development and Use of AI/ML in Biomedical and Behavioral Sciences” on February 3, 2022. The goal of this notice was to make the data generated through NIH-funded research AI/ML-ready and shared through repositories, knowledgebases, or other data sharing resources.
Meetings and Reports
Closed Funding Opportunities:
- 2022: (NOT-OD-22-065) Expired April 1, 2022. Frequently Asked Questions (FAQs)
Twenty-two awards were made in 2022 to principal investigators at 33 different institutions across the country. Awardee projects and their descriptions are available below.
- Open all
- Close all
| Principal Investigator | Institution | Project Title | NIH IC |
|---|---|---|---|
| Bui, Alex | University of California Los Angeles | PREMIERE: A PREdictive Model Index and Exchange REpositoryDeveloping novel artificial intelligence algorithms to accurately measure neurodegeneration in laboratory models of Alzheimer’s disease and related dementias, eliminating common human bias in image analysis. | NIBIB |
| Disis, Mary L | University of Washington | Developing Community-Responsive mHealth and AI/ML: Understanding Perspectives of Hispanic Community Members in Washington StateThe purpose of this project is to identify and develop solutions to ethical challenges that could impede adoption of artificial intelligence technologies for cognitive impairment screening in primary care practices through interviews with clinicians and patients and a systematic review of the existing literature. | NCATS |
| Do, Richard Kinh Gian | Sloan-Kettering Inst Can Research | Development and Validation of Prognostic Radiomic Markers of Response and Recurrence for Patients with Colorectal Liver MetastasesEnsuring equitable AI in healthcare by embedding fairness into machine learning model optimization with real-world data. | NCI |
| Federman, Alex D | Icahn School of Medicine at Mount Sinai | Natural Language Processing and Automated Speech Recognition to Identify Older Adults with Cognitive Impairment SupplementTesting the effectiveness of explainable AI in enhancing stakeholder trust and detecting bias in AI models. | NIA |
| Finkbeiner, Steven M | J. David Gladstone Institutes | Cell and Network Disruptions and Associated Pathogenenesis in Tauopathy and Down SyndromeWe use interactive Virtual Reality technology to create educational materials to promote the knowledge and practice of ethical AI among medical AI developers | NIA |
| Goldstein, Benjamin Alan | Duke University | Predictive Analytics in Hemodialysis: Enabling Precision Care for Patient with ESKDDeveloping, refining, and pilot-testing an ethical framework-guided metric tool for assessing bias in Big Data studies using EHR datasets through interdisciplinary dialogues, in-depth interviews of key stakeholders of Big Data research, and community charette workshops to gather input from ethics experts, disciplinary experts, clinicians, data scientists, and patient representatives. | NIDDK |
| Herrington, John David | Children's Hospital of Philadelphia | Ethical Perspectives Towards Using Smart Contracts for Patient Consent and Data Protection of Digital Phenotype Data in Machine Learning EnvironmentsUnravelling AI bias in colorectal cancer algorithms, identifying meaningful biomarkers, and providing guidance for the ethical development of AI, ensuring that Black patients are not harmed by AI-based models. | NIMH |
| Holder, Andre L | Emory University | Characterizing patients at risk for sepsis through Big Data (Supplement)How do we use a mortality clinical predictive model in an ethical and responsible way? | NIGMS |
| Jha, Abhinav K | Washington University | A framework to quantify and incorporate uncertainty for ethical application of AI-based quantitative imaging in clinical decision makingDeveloping best practices and ways to understand and formally document the ethical, legal, and social issues (ELSI) and ethical AI (ETAI) considerations surrounding the use of syn-thetic datasets in predictive models. | NIBIB |
| Jiang, Xiaoqian | University Of Texas Health Science Center Houston | Finding combinatorial drug repositioning therapy for Alzheimers disease and related dementiasThis project assesses patient attitudes and beliefs about key biomedical and public health ethical principles and issues such as trust, autonomy, harm, equity, and assurance, as they relate to the expected benefit of and comfort with the use of AI/ML in radiation oncology. | NIA |
| Kamaleswaran, Rishikesan | Emory University | EQuitable, Uniform and Intelligent Time-based conformal Inference (EQUITI) FrameworkWe propose to develop methods for ethical application of artificial intelligence (AI)-based quantitative imaging tools for clinical decision making. | NIGMS |
| Langlotz, Curtis P | Stanford University | Population-level Pulmonary Embolism Outcome Prediction with Imaging and Clinical Data: A Multi-Center StudyThe goal is to develop a generalizable approach for identifying and addressing ethical challenges with AI and a roadmap for how to address ethical concerns with clinical trials of AI. | NHLBI |
| Naidech, Andrew M | Northwestern University at Chicago | Hemostasis, Hematoma Expansion, and Outcomes After Intracerebral HemorrhageThis work will use AI-based nosocomial sepsis prediction models to create an alert that detects biases in those predictions, and will compare this to the perceived biases that key stakeholders might expect. | NINDS |
| Odero-Marah, Valerie | Morgan State University | RCMI@Morgan: Center for Urban Health Disparities Research and InnovationA framework for quantifying the likelihood of error and uncertainty within time series to enable equitable and uniform predictions by artificial intelligence and machine learning algorithms. | NIMHD |
| Ohno-Machado, Lucila | University of California, San Diego | Genetic & Social Determinants of Health: Center for Admixture Science and TechnologyThe study explores the presence of disability biasing language in medical notes of blind and nonblind diabetes patients and engages in discussion on disability ethics and equity of AI/ML among clinicians, data scientists, blind adults, and ELSI researchers. | NHGRI |
| Olatosi, Bankole | University of South Carolina at Columbia | An ethical framework-guided metric tool for assessing bias in EHR-based Big Data studiesThe project combines multi-resolution haplotyping for genome-wide association studies, which does not rely on ancestry constructs, with an analysis of existing literature to understand how the adoption of such methods might meaningfully address the lack of diversity in genomic datasets, thereby promoting the ethical use of NIH-supported AI-research products. | NIAID |
| Platt, Jodyn Elizabeth | University of Michigan at Ann Arbor | Public trust of artificial intelligence in the precision CDS health ecosystem - Administrative SupplementDeveloping a generalizable ethical analysis methodology to: 1) identify the ethical issues that may emerge with individual healthcare AI applications; 2) create consensus on how to address identified issues; and 3) produce a report to communicate the identified ethical issues and possible solutions to all users. | NIBIB |
| Sabatello, Maya | Columbia University Health Sciences | Blind/Disability and Intersectional Biases in E-Health Records (EHRs) of Diabetes Patients: Building a Dialogue on Equity of AI/ML Models in Clinical CareThis project aims to understand bias in AI tools used for diagnosing patients with breathing problems, by investigating disparities in diagnostic testing and studying how these disparities impact model performance. | NHGRI |
| Sjoding, Michael William | University of Michigan at Ann Arbor | Human-AI Collaborations to Improve Accuracy and Mitigate Bias in Acute Dyspnea DiagnosisExploring machine learning techniques to reduce bias in data analysis in health disparities research. | NHLBI |
| Wolf, Risa Michelle | Johns Hopkins University | Autonomous AI to mitigate disparities for diabetic retinopathy screening in youth during and after COVID-19This study explores ethical issues, incentives and practical challenges of using blockchain-based smart contracts to automate patient-centered preferences for secondary uses of digital personal health information, and to promote transparency and patient engagement in data sharing decisions. | NEI |
| Wun, Theodore | University of California at Davis | UC Davis Clinical and Translational Science CenterHow do you feel about artificial intelligence deciding your treatment for stroke? | NCATS |
| Zeng, Qing | George Washington University | Use Explainable AI to Improve the Trust of and Detect the Bias of AI ModelsDeveloping a maturity model framework for assessing institutional and community capabilities of including ethical principles in science research development | NIA |
| Zhi, Degui | University Of Texas Health Science Center Houston | Genetics of deep-learning-derived neuroimaging endophenotypes for Alzheimer's Disease (Parent grant)Incorporating Hispanic and Latinx community members’ values and priorities into the development of AI/ML-enabled mobile health applications | NIA |