Got lots of interviews or focus groups to transcribe?
AI tools can make the job a lot easier, but they can also be problematic. What are they doing with your data? How accurate are they? What are the privacy and security provisions?
While any use of AI tools in research requires careful consideration, there are two transcription tools that AUT has recently deemed ‘very high trust’. Vibe and Buzz are both open source tools powered by OpenAI’s Whisper, and AUT research students can now download the approved desktop versions for free.
Vibe and Buzz are both very similar – which one you use will be largely a matter of personal preference. Both can produce transcriptions of live audio via your device’s microphone, of pre-recorded audio/video files, and of publicly available audio/video (e.g. YouTube videos). They are straightforward to use, and you can get a transcript in just a few clicks.
In our tests transcribing from recordings, these tools did take some time to process the audio (depending on your file type, allow about 25-35 seconds per minute of audio). However, that’s much faster than transcribing manually – and the quality was so good that only a few edits were required.
If you would like to use either or both of these tools, you can check out our guide below.
Be sure to discuss your use of these tools with your supervisor, and ensure that any use is aligned with your AUTEC approval and your participants’ consent. It’s also advisable to manually check your transcripts for accuracy as these tools can and do misinterpret audio at times.
