Terms of Service

1. Nature of the Service

The Service is a probabilistic screening tool designed for educational and research purposes. It analyzes text using statistical models to estimate the likelihood of AI generation.

2. Accuracy & False Positives

While our models are trained on extensive datasets, false positives (human text flagged as AI) and false negatives (AI text identified as human) can occur. Factors influencing accuracy include:

• Non-native English writing styles.
• Use of formulaic or technical language (e.g., legal, medical).
• Short text samples (under 50 words).
• Heavy use of editing tools like Grammarly.

The Originality Research Lab accepts no liability for errors in detection or the consequences thereof.

3. Data Privacy & Zero-Log Policy

We adhere to a strict Zero-Log Policy for public scans.

• Ephemeral Processing: Text submitted via the public scanner is processed in Random Access Memory (RAM) and is permanently discarded immediately after the session ends.
• No Training: We do not use user submissions to train, fine-tune, or improve our AI models.
• Institutional API: Data retention for institutional API users is governed by separate Data Processing Agreements (DPA) specific to each university.

4. Limitation of Liability

To the maximum extent permitted by law, the Originality Research Lab, its researchers, and affiliates shall not be liable for any indirect, incidental, special, consequential, or punitive damages, including without limitation:

• Loss of academic standing, grades, or enrollment.
• Reputational damage resulting from the interpretation of reports.
• Loss of data or profits.

5. Fair Use & Restrictions

You agree not to misuse the Service. Prohibited actions include: reverse-engineering the detection algorithms, automating scans without an API license (“scraping”), or using the Service to develop competing forensic tools.