Privacy Policy

MilkSpoilageClassifier CustomGPT

Last updated: 01/24/2026

1. Overview

This Privacy Policy explains how information is handled when users interact with the Milk Spoilage Classification CustomGPT (the “Service”), including its public web interface hosted at
https://huggingface.co/spaces/chenhaoq87/MilkSpoilageClassifier-Demo.

The Service provides research-oriented, decision-support outputs by running a machine-learning model that classifies fluid milk spoilage types based solely on microbiological count data.

2. Information You Provide

The Service processes only microbiological count data voluntarily submitted by users, including:

  • Standard plate counts (SPC)
  • Total Gram-negative counts (TGN)

No other categories of data are required or processed.

The Service does not collect or request personal data, including names, contact information, identifiers, facility addresses, or geolocation data.

3. Commercially Sensitive Information

Microbiological count data may be commercially sensitive and may reflect product quality or operational conditions. Users are responsible for ensuring that submission of such data complies with their internal policies, contractual obligations, and applicable laws.

4. Data Retention and Storage

  • The Service does not retain, store, or persist user-provided microbiological data or model outputs.
  • All data are processed ephemerally for the sole purpose of generating the requested classification during the active session.
  • No session logs, databases, or datasets containing user inputs are maintained by the Service operator.

5. Use of Third-Party Services

To generate predictions, the Service calls the Hugging Face Inference API to execute the underlying model.

Submitted data may be transmitted to Hugging Face infrastructure for inference purposes. Hugging Face’s handling of data is governed by its own terms and privacy policies. No additional third-party services, analytics tools, or tracking mechanisms are used.

6. Use of Data for Model Training

User-submitted data and outputs generated by the Service are not used for:

  • Model training or retraining
  • Fine-tuning
  • Performance monitoring or benchmarking

7. Purpose and Limitations

The Service is intended for:

  • Research use
  • Industry preview
  • Exploratory decision support

Outputs are not regulatory determinations, compliance decisions, or safety certifications, and they do not replace laboratory testing, expert judgment, or regulatory review.

8. Data Security

Reasonable technical measures are used to limit exposure of data during transmission and processing. However, no online service can guarantee absolute security, and users submit data at their own discretion.

9. Regulatory Alignment

Although the Service does not process personal data, it follows general data-minimization principles consistent with common privacy frameworks, including GDPR and applicable U.S. state privacy laws.

10. Changes to This Policy

This Privacy Policy may be updated periodically. Any changes will be reflected by an updated “Last updated” date.

11. Contact

For questions regarding this Privacy Policy or the Service, please contact:
Luke Qian, cq87 [at] cornell.edu