Study reveals granular consent paradox in cardiac AI data sharing
Older adults want granular control over health data sharing but fear complexity
A team of researchers led by Pavithren V S Pakianathan published a study on arXiv examining how older adults with chronic conditions navigate granular consent in health data sharing for AI applications. The work, presented at MuC'26 as a Work in Progress submission, involved a two-stage evaluation process combining expert workshops with user testing.
The study compared single-step versus multi-step granular consent mechanisms with cardiac patients (n=7), finding no measurable differences in usability metrics. However, qualitative insights revealed a 'control-burden paradox' where participants demanded greater transparency and control over AI processing despite the additional complexity. The research highlights the need for trust-dependent design in health data systems, suggesting future consent mechanisms should prioritize clarity over granularity to balance user control with practical usability.
- 7 cardiac patients tested single-step vs multi-step consent mechanisms with no significant usability differences detected
- Participants exhibited a 'control-burden paradox' prioritizing transparency over simplicity in AI health data systems
- Study authors include Pavithren V S Pakianathan, Rania Islambouli, and others from arXiv:2608.03533
Why It Matters
Critical insights for designing ethically sound AI health systems balancing user autonomy with practical usability