Skip to main content
IT Service Status
IT Service Status

AI Transcription of Research Interviews

Northwestern researchers have several options for transcribing research interviews securely with AI services.

Transcription services not listed below, such as Rev or Otter.ai, should not be used with Northwestern research data. Panopto, a Northwestern service, should not be used with non-public research data.

The accuracy of AI transcription services varies. It is always a good idea to test the service you plan to use with a recording similar to your research data to determine whether the accuracy level will be acceptable for your research needs.

Recording and Transcribing Together

To both record and transcribe audio or video research interviews, use Northwestern Zoom or Microsoft Teams. Recordings created with Zoom or Teams can also be processed later with other applications or services.

Zoom

  • Ensure you are logged in to your Northwestern account.
  • Record to your computer, not the cloud, in order to securely store recorded files and prevent them from being backed up to Panopto.
  • Check that the location where meeting recordings are stored is only shared with approved research team members.
  • Note: When saving a Zoom recording to your computer, a transcription is not automatically generated. To create a transcript, you must turn on transcription or live captioning after the meeting starts and save the transcript before the meeting ends. The transcript will not be saved automatically, and you will not be prompted to save it before ending the meeting.
  • Appropriate for Level 2 and most Level 3 data (see Data Classification Guidelines).
  • Should not be used with data subject to HIPAA or data use agreements that prohibit use of Zoom or AI services. If working with HIPAA data, request access instead to Northwestern HIPAA-Compliant Zoom.

Microsoft Teams

  • Ensure you are logged in to your Northwestern account.
  • Start recording. A live transcript is automatically started when you start a recording.
  • Teams recordings are saved to OneDrive. The transcript can be downloaded separately from the video during the call, or, once the call ends, from the chat instance in Teams created for the call or the recording file saved in your OneDrive.
  • Appropriate for Level 2 and most Level 3 data (see Data Classification Guidelines).

Transcribing Existing Recordings on Your Computer

Several software programs exist for transcribing recordings locally on your computer, without files or data leaving your computer. Many of these applications are built on OpenAI’s Whisper model. The applications vary in cost, and availability varies across operating systems.

When selecting a software application to use, review the security and privacy information provided with the application. Turn off all internet connections from your device to check that the software is working locally on your computer and not sending any data to the cloud.

The performance of locally-running applications may depend on the hardware specifications of your computer. Not all laptops can successfully run high-quality transcription software.

Do not use AI transcription software that sends data to third-party services or requires you to provide an API key in order to use the software.

Transcribing Existing Recordings with Cloud Services

Microsoft Azure includes several AI speech-to-text services that can securely transcribe existing recordings within a Northwestern-affiliated Azure subscription. To work with such services securely, ensure that appropriate security controls are enabled for Azure resources and only use approved AI services covered by Northwestern’s Microsoft agreements.

The cost of Microsoft Azure AI transcription services depends on the service chosen and the number and length of files processed. Transaction services relevant to processing research recordings are generally affordable.

To test these services for your research project, or get started with a Microsoft Azure account, please contact researchdata@northwestern.edu.