AI Data, Training & Model Failure Mode Research Volunteer — Healthcare AI (Remote)

Already, 1 in 10 patients are harmed during their medical care. Our 501c3 nonprofit works to ensure AI improves patient care rather than putting it at greater risk.

We host a variety of remote volunteer opportunities. From researching the use of AI in diagnostics to creating patient education tools, and from liaising with industry to investigative journalism - BRITE Institute offers diverse ways to use your skills to save lives.

,

About The Initiative

The BRITE Institute is expanding a structured AI Risk Framework initiative focused on building a foundational risk architecture for AI in healthcare.

As artificial intelligence becomes increasingly embedded in clinical decision-making, diagnostics, health data systems, patient monitoring, life sciences, regulatory workflows, and healthcare operations, there is a growing need to identify where AI systems may fail — not only technically, but also from a governance, compliance, oversight, accountability, and regulatory perspective.

This initiative is designed to identify, analyze, and document AI failure modes that may impact patient safety, regulatory readiness, operational integrity, institutional trust, and responsible AI adoption across healthcare and other high-risk environments.

This is a unique opportunity to contribute to meaningful work at the intersection of

* healthcare data

* AI/ML

* public health

* biostatistics

* health informatics

* clinical research

* data quality

* model training

* and patient safety

The project operates through a standardized research workflow and collaborative system.

Volunteer Roles Available

We are seeking contributors with interest or experience in areas such as

* health data analytics

* public health

* epidemiology

* biostatistics

* health informatics

* clinical research

* data quality

* AI/ML research

* data science

* registry data

* patient-reported outcomes

* health system analytics

* population health

* research methods

* SAS, R, Python, SQL, or statistical analysis

* clinical datasets

* quality reporting

* health equity analytics

* data governance

This is a fully remote and unpaid volunteer opportunity.

What You Will Do

Contributors may assist with

* identifying healthcare AI data and training failure modes

* analyzing how poor data quality affects AI performance and safety

* evaluating risks related to biased, incomplete, or nonrepresentative datasets

* identifying labeling, annotation, and classification errors

* reviewing risks related to missing data, outdated data, and fragmented health records

* analyzing how model training choices may create unsafe outputs

* supporting structured research documentation

* reviewing healthcare use cases involving predictive models, decision support, triage tools, diagnostics, and operational AI

* contributing to mitigation strategy development

* supporting publication-oriented research outputs

* helping translate technical data risks into healthcare safety implications

Ideal Candidate Profile

We are looking for contributors who are

* detail-oriented

* analytical

* comfortable working with research concepts

* able to follow structured workflows

* organized and reliable

* comfortable reviewing technical or semi-technical material

* able to document findings clearly

* interested in healthcare AI safety

* able to think critically about data limitations, bias, and real-world use cases

* comfortable working independently and collaboratively

Strong candidates may have experience or interest in

* health data analysis

* epidemiology

* biostatistics

* public health research

* clinical research

* health informatics

* data quality improvement

* patient registry data

* predictive modeling

* AI/ML concepts

* research methodology

* SAS, R, Python, SQL, or data visualization

* health equity or population health

* healthcare operations data

Prior AI experience is helpful but not required. Clinical experience is helpful but not required. Technical experience is helpful but not required. The strongest candidates will be able to connect AI data and training issues to real-world AI healthcare risk.

Important Notes

This initiative is fast-moving, systems-oriented, and research intensive. Contributors should be comfortable

* learning new workflows quickly

* working within standardized templates

* receiving structured feedback

* reviewing research and technical material

* meeting deadlines

* documenting findings consistently

* using collaborative digital tools and remote workflows

* Remote collaboration may include Slack, Zoom, Google Meet, Google Docs, shared spreadsheets, templates, and structured research data entry systems.

Because this work may contribute to future publications, policy frameworks, and advanced AI safety initiatives, professionalism, accuracy, reliability, and attention to detail are extremely important.

What You’ll Gain

Volunteers may gain

* exposure to emerging AI safety research

* experience analyzing healthcare AI data risks

* publication-oriented collaboration

* experience contributing to a structured AI risk framework

* hands-on exposure to AI/ML issues affecting healthcare systems

* interdisciplinary experience across healthcare, data, and technology

* systems-thinking experience in patient safety and responsible AI deployment

* opportunities to help shape safer, more reliable, and more equitable healthcare AI systems

This is an opportunity to help identify the hidden data and training weaknesses that can determine whether healthcare AI systems improve care — or create new risks.

This is a merit-based and experienced-based volunteer opportunity.

Applicants should be prepared to submit relevant work samples, research examples, writing samples, regulatory experience, policy work, or related professional experience during the application process.

Selected candidates may also be asked to complete a short skills-based assessment aligned with the responsibilities outlined in this position description.

Remote

* Volunteer

* Unpaid

* Flexible Hours