How CloudCard handles data used for machine learning models

How CloudCard Uses Machine Learning for High-Quality ID Photos

At CloudCard, we use custom-built machine learning models to ensure that the ID photos provided to our customers meet high standards of quality.

How We Use Data to Improve Our Models

To train these models effectively, we utilize de-identified photo data that cardholders submit. "De-identified" means that any identifying information is removed well before these photos are used in training. This data helps our algorithms learn and improve, but no personal information—such as names, demographic details, or other personally identifiable information (PII)—is ever included. Even photo metadata, like timestamps or location data, is stripped from the photos before they are used. This ensures that our models focus only on what's important for photo quality without learning unintended patterns.

Data Security and Bias Prevention

We train all of our machine learning models in-house, ensuring that we have full oversight over how your data is used. We do not work with any third parties in this process. The models themselves never actually "see" any customer-specific information or retain details about individual photos. This setup allows us to prevent biases and ensures the models aren’t learning patterns that could impact fairness or security.

What Happens to Your Data?

The finished model is essentially a deep neural network designed to recognize high-quality ID photos. Think of it like a well-trained bank teller who can tell the difference between real and counterfeit money: the teller retains the ability to spot counterfeit bills without needing to keep any examples. Similarly, our models learn from the photos submitted without holding onto any actual customer data.

Your Choice to Participate

We also respect our customers’ preferences around data usage. You are always welcome to opt out if you’d prefer that your data not be included in our model training process. We’re committed to delivering the best quality and security in a way that aligns with our customers’ comfort and control over their own data.

Visit our Trust Center

Want more information about this matter? Please visit our Trust Center at trust.cloudcard.us or click here.