Samsung Health users are being met with a new in-app demand: agree to let Samsung use personal health records for AI training and modelling, or lose access to cloud synchronization—and watch already-synced data get wiped from Samsung’s servers. The ultimatum, first spotted by How-To Geek on July 10, 2025, appears as a toggle labelled “Consent to the Use of Health Data for AI Training and Modelling.” Flipping it off doesn’t just opt you out of AI processing. It stops all Samsung Health syncing to the Samsung account and deletes the data previously stored there, unless Samsung says it must hang onto it for legal reasons.

It’s a jarring shift from the way most consumer apps handle AI consent. Typically, you can reject a new AI-powered feature and still use the core service. Here, declining the AI side of things breaks a fundamental function that users have relied on for years—cloud backup and cross-device syncing. And for Windows users who keep their Galaxy phones linked via Phone Link or who review health dashboards on a PC, that cloud sync is often the only bridge to their data when devices are replaced or reset.

The prompt itself is clear about the consequences. Samsung’s warning text explains that withdrawing or refusing consent “stops Samsung Health from syncing health data with the user’s Samsung account” and triggers deletion of that account-linked data. It’s not a threat to wipe every bit of data instantly from your phone; locally stored records on the device itself should remain untouched. But for anyone who has ever relied on Samsung’s cloud to carry history from an old phone to a new one, the message is sobering.

That cloud deletion matters most when you’re upgrading your Galaxy phone, reinstalling the app after a factory reset, or trying to access your health timeline on a Windows machine through Samsung Health’s web portal or the Phone Link app. Without synced data, years of steps, workouts, sleep analyses, and body composition scans vanish from the cloud. And Samsung’s own support documentation warns that intentionally erased data cannot be recovered.

What’s at Stake: A Health Data Trove

The data Samsung is asking to train on goes well beyond step counts. According to Samsung’s public support pages, a typical Samsung Health account can hold:

  • Daily activity, exercise routes, step totals, nutrition logs, weight trends, and body composition readings.
  • Continuous heart rate, blood oxygen, skin temperature, and stress measurements from Galaxy watches and rings.
  • Manually entered medications, dosage schedules, and prescription details.
  • Full Health Records—conditions, immunizations, allergies, diagnoses, and test results imported from healthcare providers.
  • Menstrual cycle tracking, including period dates, symptoms, moods, and fertility estimates.

Stitched together, that data paints an astonishingly detailed physiological and behavioral portrait. Samsung’s Consumer Health Data Privacy Statement, updated in July 2025, already defines consumer health data broadly, covering vital signs, reproductive information, medical records, and even inferences derived from algorithms. The new consent request explicitly includes all of that for AI training—“to enhance health analysis, among other AI functions,” as the company puts it.

Notably, Samsung says the training may involve human review. And we don’t yet know whether the data is anonymized or pseudonymized before it’s fed into models, how long it’s retained, or whether already-trained models can be compelled to forget data after consent is pulled back. Those unanswered questions are magnified by the sensitivity of the information being handed over.

How We Arrived at This Ultimatum

Samsung has been steadily weaving AI into its health ecosystem. In June 2025, the company announced deeper analysis features for Galaxy watches and rings: cardiovascular load assessments, fitness comparisons against peers, and proactive wellness tips, all driven by accumulated user data. Building those models requires huge and varied datasets—exactly what Samsung Health’s billions of data points could provide.

But the business logic collides with user expectations around consent. Health data sits in a special category, legally and emotionally. Meaningful consent under privacy regulations like GDPR or certain state laws requires a choice that is freely given and not bundled with unrelated services. Tying refusal of AI training to the loss of cloud sync—a feature users have depended on for years—risks making that consent feel compulsory.

The company hasn’t said whether the rollout is global or limited to specific regions. Language in the prompt and its consequences might vary by jurisdiction. Reports indicate that only some users are seeing the toggle so far, which suggests a phased or A/B testing approach. But the absence of a clear, widely published explanation means uncertainty is rampant.

What This Means for You

If you’re a casual Samsung Health user who primarily checks step counts and doesn’t rely on cloud to move data between devices, the immediate impact is smaller. You can probably refuse consent and continue using the app’s local features just fine—though Samsung may limit future AI-driven insights that build on your history.

If you’ve been logging years of workouts, sleep cycles, medication schedules, or cycle tracking, and you depend on Samsung Health syncing to keep that data safe across phone upgrades, the stakes are higher. Declining AI consent could erase exactly the backup you depend on. And for anyone using a Windows PC to view their health dashboard—perhaps through Samsung Health’s web interface or the Your Phone app—a wiped cloud means your PC no longer sees recent records.

IT administrators managing Samsung devices in a workplace won’t see a direct enterprise tool here, but the principle matters. Employees who store health data on personal devices may not realize that a consumer consent screen could nuke years of personal records. And from a data governance perspective, the lack of clarity on anonymization and retention should raise red flags for any organization that advises users on health privacy.

How to Safeguard Your Records Before Deciding

Before you tap “agree” or “disagree,” grab a copy of your data. Samsung Health has a built-in export tool that saves your history into files you can store elsewhere.

  1. Open Samsung Health and tap the three-dot menu.
  2. Go to Settings.
  3. Select Download personal data.
  4. The app will generate a zip file and place it in a Samsung Health folder accessible through your phone’s file manager.

Once the export is complete, connect your phone to a Windows PC via USB. Copy that folder to an encrypted location—a BitLocker-protected drive on Windows, for example—so you control the only copy. Even if you do consent to AI training, having a manual backup is never a bad idea.

If you use Health Connect on Android, some of your health records might already be shared with other apps. But Health Connect isn’t a comprehensive backup of Samsung Health, and it doesn’t capture Samsung-specific metrics like stress or body composition scores. Treat the manual export as your primary safety net.

And don’t confuse this new AI toggle with Samsung Health’s general data processing consent. The app’s privacy settings page may now display multiple switches; accidentally flipping the wrong one could disrupt ordinary syncing before you’ve secured your data.

The Road Ahead

Samsung hasn’t yet explained why AI consent and cloud sync need to be tied together. A reasonable user might expect to reject AI training and continue syncing encrypted health data to their account, just as they’ve always done. The company’s silence on anonymization, retention periods, and the fate of models trained on withdrawn data only deepens the trust gap.

Regulatory scrutiny could be next. Health data is tightly regulated in many regions, and opt-outs that come with punitive cloud deletions may attract attention from data protection authorities. Samsung would be wise to clarify whether users who initially consent but later withdraw will have their data retroactively removed from training sets, and how they can verify that.

For now, the single most important step is to export your health history before touching that toggle. Samsung’s dialog offers a choice; preparing a full backup ensures that choice doesn’t become a regrettable one.

Microsoft doesn’t have a direct equivalent to Samsung Health, but the episode is a lesson for any cloud-connected health platform. As AI moves deeper into personal wellness, the line between feature and coercion must remain unmistakable—and users should never have to trade their own history for the right to keep it safe.