Whisper is an open-source speech-to-text tool developed by OpenAI that can be used to transcribe audio and video files. By using HPC-UGent (High Performance Computing), you can easily and quickly generate high-quality transcriptions of interviews, in multiple languages. Furthermore, you data stays safe, because the model is run locally.

Why use Whisper?

• Whisper generates high-quality transcriptions in different languages, and can translate these transcriptions simultaneously.
• Using HPC-UGent’s computing power, transcriptions are generated fast, saving you a lot of time on manually transcribing interviews.
• Privacy: as interviews often contain personal data that should not be shared, data privacy is an important issue. HPC-UGent is a service ran on local UGent hardware (servers). This ensures that files transcribed by Whisper on HPC-UGent remain stored locally and are not shared externally with a cloud service.

How can you use Whisper?

• Before you are able to connect to HPC-UGent, you need to create an account. With an account, you can connect to the HPC web interface.
• In the interface, you upload the files you want to transcribe to your home directory under ‘Files’.
• You can find Whisper under ‘Transcribe’ in the ‘Interactive Apps’ dropdown.
• You can then select the input file and original language. If you want to translate the transcription, select a translation target language. You don’t have to change the rest of the settings.
• By clicking ‘Launch’, Whisper will start to transcribe the file.
• Once the task is done, you can find the transcription in your home directory.

For more information, you can watch this short webinar (in Dutch, with English subtitles).

Need more information about (Gen)AI?

Quite a lot of information regarding the use of (generative) AI, from different perspectives and with different objectives, is already available at Ghent University.

More information regarding (Gen)AI and research?

• Intranet: Artificial Intelligence and Research
• Research Tip: GenAI: responsible use in research
• Research Tip: GDPR: What should I take into account when developing or using AI? 
• Research Tip: GenAI: Overview of National and European legislation and guidelines
• Research Tip: GenAI: Writing a scientific article
• Research Tip: GenAI: Writing a (postdoctoral) research proposal
• Research Tip: GenAI: PhD research (proposal) and AI
• Research Tip: GenAI: Transcription of interviews with Whisper
• Mind the Gap Podcast: Use of Gen(AI) in research: here to stay (episode 4)
• Mind the Gap Trainingtool VLIR training on good academic research (see “Extra - Ethics rules and regulations” )

More information regarding (Gen)AI and education?

• Education Tip: Generative AI in Ghent University Education: Impact and Approach

Do you want to learn, experiment and practice?

• Experimentation Guidelines: How to use the AI Tools
• Ufora Infosite: Generative AI for teachers: Concepts, Creations and Classroom Practice (also relevant for researchers!)
• Courses and info session: please consult the UTOP-platform
• Get inspired by other UGent'ers in the monthly online exchange sessions AI-nspiratie over de middag: hoe gebruiken UGent’ers AI? (sessions alternately in Dutch and English).

Do you want to know to which information the students have access?

• Info for Ghent University students: Generative AI in Education at Ghent University