Issue Medical-act definitions, data ownership and interoperability gaps for hyper-personalised health platforms

Summary

If the AI in a hyper-personalised health platform detects anomalies and warns patients directly or adjusts their medication, the function could be deemed a medical act that requires device approval from Korea's drug safety regulator. Combining clinical, genomic and lifelog data for multi-omics analysis runs into unclear data ownership and strict standards for handling pseudonymised information, which restrict integration. Missing standards such as HL7 FHIR between wearables and hospital electronic medical records fragment data and limit analysis. When AI judgements lack medical explanation, doctors and patients find them hard to trust, which slows adoption in practice. Digital products also face high user drop-out and a digital placebo effect, so genuine algorithm effects must be separated from the effect of use itself.