The Human Values Project

Clarifying Whom AI Serves in Healthcare

Investigating how to explicitly encode ethical, cultural, and personal values into AI-driven clinical decision-support systems so that recommendations align with patient, clinician or societal norms.

The Challenge

As AI systems become more prevalent in healthcare, a critical question emerges: How do we ensure that AI recommendations align with human values?

Different patients, cultures, and communities may have varying perspectives on treatment preferences, end-of-life care, risk tolerance, and quality of life considerations.

Key Questions We Address:

  • How can AI systems understand patient preferences beyond clinical data?
  • What frameworks exist for incorporating cultural and ethical considerations?
  • How do we balance personalization with evidence-based medicine?

Our Approach

  • Value/preference elicitation based on techniques pioneered in cognitive science.
  • Culturally localized frameworks
  • Patient preference modeling
  • Ethical AI design principles

Research Components

Multi-faceted approach to value-aligned AI in healthcare

Value Elicitation

Developing methods to systematically capture and represent patient values, preferences, and cultural considerations in computational formats.

  • • Patient preference surveys and interviews
  • • Cultural value mapping
  • • Preference learning algorithms

Ethical Decision Frameworks

Creating structured frameworks that integrate ethical principles with clinical evidence to guide AI-assisted medical decisions.

  • • Multi-criteria decision analysis
  • • Ethical principle integration
  • • Value-based decision trees

Cultural Adaptation

Ensuring AI systems can adapt their recommendations based on cultural backgrounds, religious beliefs, and community norms.

  • • Cross-cultural healthcare studies
  • • Religious and spiritual considerations
  • • Community-based preference modeling

Transparency & Trust

Building transparent AI systems that can explain how values influence their recommendations and maintaining trust through interpretability.

  • • Explainable AI for value-based decisions
  • • Trust measurement frameworks
  • • Recommendation transparency tools

Clinical Implementation

Translating value-based AI frameworks into practical clinical tools that can be integrated into existing healthcare workflows.

  • • Clinical decision support integration
  • • Workflow optimization
  • • Real-world validation studies

Policy & Governance

Developing governance frameworks and policy recommendations for ethical AI deployment in healthcare systems.

  • • Ethical AI governance models
  • • Regulatory framework development
  • • Stakeholder engagement strategies

Real-World Applications

How value-based AI transforms clinical decision making

End-of-Life Care Decisions

Developing AI systems that consider patient values regarding quality of life, comfort care preferences, and family involvement in end-of-life treatment decisions.

Incorporating advance directive preferences
Balancing aggressive treatment vs. comfort care
Cultural and religious considerations

Treatment Personalization

AI systems that consider individual patient values such as risk tolerance, lifestyle preferences, and personal goals when recommending treatment options.

Risk tolerance assessment and integration
Lifestyle impact considerations
Personal goal alignment

Key Publications

Research advancing the field of value-aligned healthcare AI

Featured Research 2024

Medical Artificial Intelligence and Human Values

Comprehensive analysis of how human values influence AI outputs in clinical settings, providing frameworks for incorporating ethical considerations into medical AI systems.

NEJM AI Ethical AI Clinical Decision Support
High Impact
Premier Medical Journal
Open Source Ongoing

Human Values Project Repository

Open-source code, data schemas, and examples of value-based decision trees. Collaborative research platform for developing ethical AI frameworks in healthcare.

GitHub Open Source Collaborative Research
Open
Community Driven

Shape the Future of AI in healthcare for autonomous humans

Join our mission to ensure AI systems in healthcare respect and incorporate human values. Together, we can build more ethical and patient-centered AI solutions.