Automatic captions & transcription
Automated captions and transcription are different things, but they use the same underlying technology called automated speech recognition (ASR). Machine captioning and transcription tools are built in to many modern operating systems, note-taking apps, and video hosting and streaming platforms. Many of these tools use AI to improve results.
Accuracy
The accuracy of ASR software varies a lot. The industry boasts accuracy rates of 95% or better, but there are many factors that affect the accuracy the machine learning algorithms. They tend to be trained on native US/UK English speakers using common or tech-related terminology. They can struggle with:
- languages other than English
- accents other than standard American and UK English
- terminology from specialized fields
- speech outdoors, with background noise, low-quality mics, or a reverberant space
Machine captions and transcripts may need to be edited to be useful. Automated captions and transcripts are a convenient and low-cost way to make events and recordings more accessible when human captioning and transcription isn't available.
OS tools
Windows live captions
Windows 11 has a built-in system for generating live captions (but not transcripts). It works with all Windows apps and can caption speech input directly from your mic. Learn more about live captions in Windows. Processing is done on your device, not the cloud.
Mac, iPhone & iPad
Many of the apps built into Apple products can generate live captions and transcriptions. These include iPhone calls, Notes for iPhone and Mac, and Voice Memos. Processing is done on your device, not the cloud.
These features were added in iOS/iPadOS 18 and macOS 15 (Sequoia). They aren't available on devices running older Apple operating systems.
Android
Most Android phones have built-in live captioning and transcription services. Details may depend on the make and manufacturer of your device. Learn about Live Caption and Live Transcribe Samsung Galaxy phones and live caption services on Pixel phones and other compatible devices.
Video meeting apps
Microsoft Teams
Licensed by UVic
Microsoft Teams can generate live captions for audio and video meetings. When you record a Teams meeting, it will also save a transcript and closed captions with it. You can read, play, edit and download the video and transcript in Microsoft Clipchamp Video Hub (formerly Stream). Teams Premium can also provide real-time text translations. Read about Microsoft Teams at UVic.
Zoom
Licensed by UVic
Zoom can automatically generate live captions and transcripts for video meetings. Captions will also be generated and saved with cloud recordings (not available to students). Read about Zoom at UVic.
Video platforms
YouTube
YouTube can produce captions, transcripts, and translations in multiple languages. Learn how to use automatic captioning in YouTube.
Privacy: At UVic, YouTube is generally recommended for publishing videos for a general audience. It's not suitable for internal, confidential or highly confidential data.
Microsoft Clipchamp Video Hub
Licensed by UVic
Clipchamp Video Hub (formerly Stream) provides AI-generated captions and transcripts for videos. Go to Clipchamp on the web or learn how to use autocaptions in Clipchamp.
Echo360
Licensed by UVic
Echo360 (a video streaming platform integrated with Brightspace) creates an interactive transcript of every video published to it. They aren't generated in real time, but are typically added shortly after the video itself is processed. The video owner can choose to make the transcript available as closed captions as well. Learn more about UVic's learning and teaching systems.
Note-taking apps
Microsoft Word and OneNote
Licensed by UVic
Word and OneNote (Windows and web apps only) can take advantage of Microsoft's cloud transcription services to transcribe voice recordings. You can use live transcription and upload pre-recorded voice recordings to transcribe. Learn how to transcribe voice recordings in Word and OneNote.
Other apps
Existing PIAs
We've already done privacy impact assessments (PIAs) on these products and found them suitable for some purposes. If you want to purchase them for your department's use, you'll need to submit it for technical approval. You may also need to develop a PIA addendum that's specific to your use case. (A PIA addendum is shorter and simpler than a full PIA.)
| Product | Use case |
|---|---|
| Ava Live Captions (Enterprise plan) | General |
| Plaud | General |
| ScribeBerry | Medical transcription |
| Heidi Health | Medical transcription |
Other apps
If you're interested in using an app not listed here for UVic work, you'll need to submit it for technical approval, even if it's free. Most ASR applications use cloud storage and processing so a PIA may also be necessary.
If your use case doesn't require you to record or process internal, confidential, or highly confidential data, we can approve most tools (with conditions). If you are working with sensitive UVic data, these approval processes help make sure private and sensitive data isn't leaked or exposed to threat actors. We look for software that has an industry standard security certification (ISO 27000 or SOC 2 Type 2) and doesn't:
- train AI systems using your data,
- sell your data to other companies,
- send unsolicited advertising to people in your contacts list, or
- claim ownership of, or a broad license to use, your data.
UVic doesn't restrict what ASR software students can use for their own research and learning on their own devices. Some students have accommodations from CAL to record lectures for personal use and may use devices and software not listed here. However, unauthorized AI agents like Otter.ai aren't permitted to join UVic Teams and Zoom meetings.
Professional transcription
Human transcription, captioning and ASL interpretation are available to D/deaf and hard of hearing students who are registered with the Centre for Accessible Learning (CAL).
Transcription, captioning and ASL interpretation for UVic events is coordinated by Equity & Human Rights (EQHR).
Privacy & security
Legally, recording a transcript of a meeting or conversation is similar to making an audio or video recording. BC's Freedom of Information and Protection of Privacy Act, FIPPA, applies to transcriptions as well. Before you record a meeting or conversation, inform everyone who could be recorded. FIPPA applies to all faculty and staff.
Get help now
Contact IT support to:
- report an outage or issue with a supported application
- consult about options for machine captioning and transcription
- get a quote for purchasing software
- get help requesting technical approval or a PIA