Privacy Statement

Privacy Statement

This Privacy Statement is not a Privacy Policy. See Brown's Privacy Policy for Brown University's official privacy policy.

In this document we describe how user contributed data to CCV AI Services, such as user-uploaded media files and user's conversations with AI models, are handled by CCV AI Services.

User data handling - Transcribe

DATA SHARING & RISK The Transcribe Service is currently permitted to handle data at Risk Level 2 and below. About Risk Classifications | View Risk Classifications for Brown Software & Services

User-submitted audio/video files

While using the Transcribe service, the users will be prompted to upload the audio/video files that they want transcribed. These files will be uploaded through a secure connection to a Google Cloud Storage (GCS) bucket managed by CCV, where all data is stored with AES-256 encryption. Only authorized personnel at Brown University will have access to the audio files. The files are stored temporarily for up to 7 days. Within this window, the users will be able to play the audio file or the audio within the video file to correct any errors in the transcripts generated. We have an automated procedure that runs daily and deletes any files in the GCS bucket that are older than 7 days.

If the OpenAI Whisper model is selected, the audio files are processed securely in secure computing resources managed by CCV. NO audio/video files are transmitted to a 3rd party provider.

If the Google Gemini model is selected, the Google Gemini Vertex AI API will require access to the audio/video file to produce the transcription. Google will NOT use your data for training purposes, and no data will be retained by Google for using this API. For more information, please see Google's Data Governance page for more information.

Transcriptions

The transcriptions from the user-submitted audio and video files are temporarily stored in a Google Cloud Firestore database managed Google Cloud. By default, Firestore encrypts the data before writing it to disk, so all data stored in Firestore is encrypted. Only authorized personnel at Brown has access to this database. The transcriptions stored in the database will not be used for any other purposes than for the users to retrieve. The users can retrieve only their own transcriptions at any time through the web interface from the associated job page. Unlike the audio/video files, the transcriptions will not be deleted unless the users choose to.

Deletion of user data

The users can choose to delete the content that they have submitted to us at any point.

Before starting a transcription job, they can delete any of the audio/video files that they have submitted to the Transcribe service via the trash can buttons the Files Uploaded table. They can also close the Start Job form or click the "Cancel" button, at which point their uploaded audio files will also be deleted.

At any point after a transcription job is completed, the users can choose to delete the job either through the Delete button in the All Jobs table on the Transcribe page or the Delete button on the View Job page. This will delete all audio files and all associated transcriptions from CCV AI Services. We will retain some metadata for the job for billing and record-keeping purposes, such as the durations of the audio files and the the timestamps of when the jobs are created and finished.

The deletion is permanent. You will no longer be able to retrieve the audio files or the associated transcriptions after deletion.

User data handling - Ask Oscar

As of April 1, 2025, we currently do not collect or store any user data for this tool. Any chat messages that the users provide and the output generated by AI models are not stored by the service. All AI models that we use for this service comes from Azure OpenAI and strictly follow the data, privacy, and security policy for Azure OpenAI, which guarantees that your prompts (inputs) and completions (outputs), your embeddings, and your training data:

• are NOT available to other customers.

• are NOT available to OpenAI.

• are NOT used to improve OpenAI models.

• are NOT used to train, retrain, or improve Azure OpenAI Service foundation models.

• are NOT used to improve any Microsoft or 3rd party products or services without your permission or instruction.

User data handling - LibreChat