May 2026 Session: Distributed Open Justice Oversight (DOJO): Open-Source Accountability for Healthcare AI-Distributed Open Justice Oversight (DOJO): Open-Source Accountability for Healthcare AI

Distributed Open Justice Oversight (DOJO): Open-Source Accountability for Healthcare AI

Video Transcription

Video Summary

The webinar introduced the AIML task force’s educational work and featured DOJO (Distributed Open Justice Oversight), a community-oriented initiative for evaluating and improving AI in healthcare.<br /><br />Speakers argued that conventional measures such as AUC can conceal serious failures for particular patient groups, especially populations underrepresented in training data. DOJO aims to offer locally installable audit tools that assess models throughout development, identify issues such as bias and shortcuts, and document how developers respond to feedback. This audit trail could help institutions judge whether a model suits their data and patients.<br /><br />The speakers emphasized that AI oversight must involve people historically harmed or excluded from healthcare—not just academics, developers, or institutions. They cautioned that frameworks and human review can become performative or ineffective unless their purpose, participants, and accountability are continually reconsidered. Accountability, they argued, should be shared across the healthcare ecosystem.<br /><br />Rather than treating AI as a technical fix, the discussion framed it as an opportunity to rethink clinical work, medical education, and relationships between clinicians and patients. The goal is not simply efficiency, but more meaningful care and systems shaped through inclusive, ongoing community collaboration.

Keywords

AIML task force

DOJO

Distributed Open Justice Oversight

AI healthcare auditing

model bias detection

AUC limitations

healthcare equity

community oversight

accountability in AI