Research Questions

• What can evidence actually prove?

• What can a clinical trial establish about an individual patient?

• When do findings transport to a new person, institution, population, or environment?

• When should medical AI advise, and when—if ever—should it decide?

• How do adaptive models change validation, approval, monitoring, and liability?

• When does prediction improve care without establishing causal understanding?

• Can AI discover scientific laws, or only fit increasingly powerful patterns?

• Which parts of clinical judgment are informational, relational, moral, or irreducibly human?

• Where can AI accelerate science and medicine with justified confidence?

• How should evidence standards change when deployment changes behavior, treatment, or the data-generating process?