Autonomous FAIR Digital Objects: From Passive Assertions to Active Knowledge

Researchers introduce Autonomous FAIR Digital Objects (aFDOs), a framework that transforms static scientific data into self-governing entities capable of validating evidence, resolving contradictions, and updating confidence independently. The system combines semantic web standards with Byzantine-fault-tolerant consensus mechanisms to enable scientific knowledge to persist and evolve beyond institutional stewardship.

This research addresses a fundamental problem in scientific infrastructure: the fragility of centralized knowledge curation. When scientific registries close or institutions cease operations, the stewardship of published data stops even though the information remains accessible online. The aFDO framework solves this by embedding autonomous governance directly into digital objects through three technical layers—policy rules aligned with semantic web standards, event announcement systems, and Byzantine-fault-tolerant consensus mechanisms. The approach treats scientific data as active entities rather than passive assertions, enabling them to evaluate incoming evidence and reconcile contradictions without human intervention. The implementation tested on 4,305 FAIR Digital Objects from rare-disease ontologies demonstrates practical viability. The consensus mechanism resolved 56.3% of naturally occurring conflicts in ClinVar submissions, with graceful degradation under adversarial conditions up to its Byzantine-tolerance threshold of f < n/5. This represents significant progress toward decentralized scientific infrastructure where knowledge can outlive its original publishers. For the broader research ecosystem, aFDOs offer a pathway toward more resilient knowledge infrastructure that reduces dependency on institutional continuity. However, the technology remains at the operational prototype stage, with questions remaining about scalability across diverse scientific domains and adoption barriers for institutions accustomed to centralized curation. The Byzantine-fault-tolerant design is particularly valuable for scientific applications where data integrity and conflict resolution matter critically.

• →aFDOs enable scientific data objects to autonomously validate evidence and resolve contradictions without institutional oversight.
• →The framework combines RDF-star, ActivityStreams, and Byzantine-fault-tolerant consensus to create portable, standards-aligned governance rules.
• →Testing on 4,305 rare-disease ontology objects achieved 56.3% conflict resolution rate matching expert panel adjudication.
• →System maintains Byzantine-fault tolerance up to f < n/5 adversarial nodes, then degrades predictably beyond that bound.
• →Architecture decouples knowledge curation from institutional continuity, addressing scientific infrastructure fragility.