Mostly True
Claims have surfaced that the Health Insurance Review & Assessment Service (HIRA) discarded valuable medical data that could have been used for AI development due to insufficient storage capacity. The allegation, however, rests in part on an explicit estimate — a figure of "hundreds of terabytes" — and there is also evidence pointing to active internal AI use at the agency, leaving key elements of the claim unconfirmed.
The core of the allegation is that HIRA, which holds one of Korea's largest collections of health insurance claims data, disposed of data because its storage capacity was insufficient. The claim characterizes the loss as significant for AI development in healthcare, where large volumes of medical data are a critical input for training and research.
The "hundreds of terabytes" figure attached to the claim is an explicit estimate rather than a confirmed measurement, and its accuracy has not been established. No verified account of the precise volume or the circumstances of the alleged disposal is available from HIRA itself.
There is also contrary evidence suggesting that HIRA's internal AI utilization is active. If the agency is actively applying AI internally, the picture of a repository of unused data being lost is less straightforward than the allegation implies. HIRA operates an open data portal that publishes insurance review datasets, which forms part of the context for the claim.
The claim that HIRA lost valuable medical data over storage constraints has a plausible basis, but the headline figure of "hundreds of terabytes" is an explicit estimate and evidence of active internal AI use cuts against the most dramatic reading. Verdict: Mostly True