Fruitly utilizes advanced artificial intelligence to provide seamless fruit identification. This section explains how AI models interact with your data, the measures taken to ensure your privacy, and the controls you have over the processing of your images.
Certain features in Fruitly depend on the processing of visual data through automated algorithms. By using the scanning feature, you allow the App to analyze image characteristics to provide accurate identification and ripeness information. AI analysis is performed strictly for informational and educational purposes.
a. Image Analysis (Required for Core Functionality)
Object Recognition: Identifies fruit varieties within the frame.
Visual Characteristic Analysis: Evaluates ripeness, color, and texture.
Encrypted Transmission: Images are secured during the identification process.
b. Model Improvement (User-Controlled)
Algorithm Training: Anonymous usage of images to improve identification accuracy.
De-identification: All metadata is stripped to ensure images cannot be linked to your account.
Opt-out Capability: This can be toggled at any time in the Settings → Consent Settings section.
a. No Biometric Data Collection
The AI models are strictly configured for botanical identification.
The Application does not perform facial recognition or any form of human identification.
Incidental captures of hu man figures are not processed and are deleted from the active cache.
b. Secure Data Handling
Third-Party AI Infrastructure: Trusted providers act as data processors and are contractually forbidden from using your images for their own commercial purposes.
Data Minimization: Only the minimum data required for a specific scan is transmitted.
a. Probabilistic Results
AI identifications are based on probability and are not guaranteed to be 100% accurate.
The Application cannot determine hidden factors like internal contamination or allergen content.
b. Final Verification
You are solely responsible for verifying the safety of any fruit before buying or consuming it.
Reliance on App results is done entirely at your own risk.