Samsung has begun requiring users to consent to sharing their health data for AI model training as a condition of maintaining full access to Samsung Health’s core features.
When opening the app, users now see a notification titled “Consent to use health data for model and AI training,” with an option to accept or decline. The data Samsung says it may draw on is broad: physical activity records, health logs, medication information, and menstrual cycle data. Some of that data may also be subject to what the company describes as “human review” as part of its model development process.
Users can refuse. But the consequences are significant. Samsung warns that declining will prevent data from syncing with their Samsung accounts, and that stored health data may be deleted unless legal requirements oblige the company to retain it.
Whether Samsung can link that data directly to individual identities remains unclear. Standard industry practice involves anonymising and aggregating data before use, making it difficult to trace back to specific people. Still, the core concern is structural: opting out of AI training effectively degrades the app’s functionality, putting users who value their privacy in a difficult position.
The move comes as Samsung expands its AI ambitions in digital health. In June, the company announced a round of Samsung Health updates built on AI, including personalised training plans, sleep analysis, and dietary recommendations. Those features are expected to arrive first on Galaxy Watch 9 devices, which Samsung is widely expected to unveil at its Galaxy Unpacked event on 22 July, before rolling out to older wearables over time.
Training AI models at that level of personalisation requires large volumes of data. Linking cloud sync to consent, rather than keeping them separate, is what makes this particular approach contentious.
