Skip to content
Partner hackathon projectApr 7, 2025Atlanta

AI Risk Management Framework for the Healthcare Sector

Vibhuti Dahiya, Maria T Mansoor · Team Maria & Vibhuti

Submitted to Georgia Tech AISI Policy Hackathon. Projects from partner hackathons are early-stage work by participants, not Apart Research publications.

Read the report

Report: AI Risk Management Framework for the Healthcare Sector

Share

This policy brief proposes an AI Risk Management Framework for the healthcare sector to address issues like data privacy, algorithmic bias, and transparency. Current federal regulations are stalled, and existing state laws and proprietary frameworks offer fragmented oversight.

The proposed framework consists of five core functions: Governance, Identification, Alignment, Management, and Continuous Improvement.

It calls on the Department of Health and Human Services (HHS) to lead its development, ensuring compliance with HIPAA and HITECH while addressing AI-specific risks.

Reviews

Judging this hackathon?

Review this project

Your public critique appears on this page without your name. Your private critique is not published; only the Apart team reads it. If you agree below, we share your review with grantmaking.ai (opens in new tab) and the Transformative AI Fund so strong projects can be funded.

Not shown on this page.

Shown on this page, without your name.

Only the Apart team reads this, and funders if you agree below.

Share my name publicly on grantmaking.ai *
Share my private critique with funders *

  1. This briefing addresses a significant problem. I would like to see the proposed framework explained in more depth (go deeper on the concerns and what you want to do) and in prose (rather than having to interpret the charts), and also see more of the "bottom line upfront" to understand right away the intended outcome. It wasn't clear to me why a framework was preferred to actual governance or policy, though. While I like this idea in general - especially because frameworks may be easier to implement than more formal regulations - I found the formatting distracting and in some cases difficult to navigate. I really enjoyed the stakeholder analysis. Overall this brief represents a good introduction to the concept of an AI risk framework for healthcare.

  2. Good overview of the state of the field with regard to AI regulation and AI in healthcare. Insightful stakeholder analysis and innovative combination of existing frameworks to develop an AI Risk Management Framework for the healthcare sector. However, it does little to contribute to advancing interpretability. One opportunity here is to provide nuanced healthcare sector information that might influence the design of AI models or data collection.

    Thorough stakeholder analysis and AI risk management framework for healthcare which have the potential for addressing AI safety concerns in the healthcare sector. Consider incorporating a section on alternative policy solutions and recommendations to improve the credibility of the plan.

    This is a well written policy brief that distills the message in a clear and concise format. Good use of infographics. However, as a policy document proffering solutions, an implementation plan would make it even stronger. As it is, there are many actions specified under the five functions that it might be daunting to know when to do what. More detailed information would enhance comprehension of the 'Impact Analysis' page.

    Read full reviewShow less

Cite this project

@misc{dahiya2025ai,
  title = {{AI Risk Management Framework for the Healthcare Sector}},
  author = {Vibhuti Dahiya and Maria T Mansoor},
  year = {2025},
  month = apr,
  note = {Submitted to Georgia Tech AISI Policy Hackathon, a partner hackathon},
  howpublished = {\url{https://apartresearch.com/sprints/projects/ai-risk-management-framework-for-the-healthcare-sector-3gol}},
  url = {https://apartresearch.com/sprints/projects/ai-risk-management-framework-for-the-healthcare-sector-3gol}
}

Build something like this at the next Sprint

AI Collusion Research Sprint · Oct 23 - 25, 2026