AI & Machine Learning - Data Use Consent
Here we kindly ask whether you would be willing to let us use your health data to train the AI that powers syncsulin? This is entirely separate from your agreement to the Privacy Policy and Terms & Conditions, and it is completely optional.
Why we’re asking
syncsulin’s goal is to give you insights that are genuinely personal. We are not aiming for generic population averages, but patterns that reflect your individual glucose behaviour. To build that kind of AI, we need real-world data from real people. Synthetic data simply doesn’t capture the variability and complexity of how glucose behaves in daily life.
Your data, combined with data from other Beta testers who also consent, is what makes the difference between a generic tracking tool and one that actually understands you. We think that’s worth being completely open about.
What data we use
- Glucose readings and trend data (CGM readings imported via Apple HealthKit or Glooko CSV)
- Manually entered insulin doses
- Meal log entries
- Sleep and activity data
- The timestamps connecting these data points — this is what makes pattern recognition possible
We do not use raw Apple HealthKit data directly for AI training. Only data that has been explicitly entered or imported into syncsulin. We do not use your name, email address, or any directly identifying information.
How we protect your data: pseudonymisation
Before your data is used for model training, it is pseudonymised: your identity is replaced with a randomly generated internal identifier. Your name, email, and the mapping between that identifier and your identity are stored separately and are never used in the training process.
Where your data goes
-
Your data is processed on EU-based infrastructure: Supabase servers in Ireland, which run on Amazon Web Services (AWS). Supabase acts as our data processor under a Data Processing Agreement (DPA) in accordance with Art. 28 GDPR.We do not sell your data. Ever.
Changing your mind
You can withdraw or grant this consent at any time in Settings → Privacy → AI Training. If you withdraw, we will stop using your data immediately for any new training. Data already incorporated into trained models cannot be retroactively removed, that’s not technically possible with how machine learning works, but no new data will be processed after withdrawal.
Legal basis: Art. 6(1)(a) GDPR (freely given, specific, informed, and unambiguous consent) and Art. 9(2)(a) GDPR (explicit consent for special category health data). This consent is legally and operationally independent from the consent required to use the app.
Consent declaration
Questions? Contact us at info@syncsulin.com