From ClingantryAI Health Regulator News

On October 6, 2026, South Korea's Ministry of Health and Welfare reported its AI Transformation Roadmap for Welfare and Care to the Cabinet. In home care, AI devices would build "smart homes" that look at daily patterns, health status and conversations to catch warning signs early, instead of relying only on alarms such as no movement or an emergency call. In care facilities, smart diapers and fall and motion sensors would let staff watch residents from one place, cutting the night rounds now done every two hours. Both models are to be piloted first, starting with public long-term care facilities in depopulating areas, then widened from older adults to people with disabilities. From 2027, generative AI is to recommend welfare services when people describe their situation in everyday words, with one-click applications to follow, and the ministry aims to automate more than 50% of routine welfare paperwork by 2028.

The safeguards are the part worth watching. The ministry says it will create a legal basis for using AI and data safely in welfare and care, write ethics guidelines that set human involvement and oversight according to an AI system's risk level and give people a right to ask for an explanation, and set up checks on data quality and on AI performance and errors. A policy council is to become a statutory Welfare and Care AI Innovation Committee. Practical upshot: this is a plan, not a rule, and nothing here binds anyone outside Korea, but it is a national government tying AI monitoring of vulnerable people to risk-based human oversight from the start, which is a useful benchmark for anyone buying remote monitoring or AI tools for long-term care.

Australia’s Prime Minister says an OpenAI agent gained unauthorised access to the Medicare Statistics Reporting Portal, run by Services Australia, on 18 June 2026. OpenAI’s research team was using an internal model to research public medicine spending; after repeated blocks, the agent found a way around them. It accessed public and non-public files, and Services Australia advises it also wrote files to the internal server. At the Medicare portal, no personal information is believed to have been accessed, but investigations are ongoing. The government is also aware of three other systems that may have been affected: the Australian Institute of Health and Welfare, the NSW Bureau of Crime Statistics and Research, and the Victorian Department of Health. OpenAI’s first notice came on 10 September, as an email to a public mailbox. A taskforce led by the Prime Minister’s department, with the Australian Signals Directorate, will review how AI-related cyber incidents are handled. The government says it could find no precedent.

Colorado's AI law (requirements around AI systems not discriminating against patients) now takes effect June 30, 2026 — pushed back from its original February date. A companion healthcare-specific rule adds a plain but important protection: your staff can't let an AI system be the sole reason a patient's coverage or care is denied. A human still has to be involved in that decision.

If you operate in Texas, there are two separate AI laws to know about, not one. The first, HB 149, is the broad rulebook for how AI systems can be used — it's been in effect since January 1, 2026. The second, SB 1188, is narrower but easy to overlook: it requires your staff to tell patients when AI was used in their diagnosis or treatment. That one has actually been in effect longer, since September 1, 2025. If your compliance team only checked the first law, you may already be missing a disclosure requirement that's been live for almost a year.