About ÈIST Team and partners Results and publications ENEnglish GDGàidhlig

Privacy & GDPR

Information about you: how we use it and with whom we share it

This tool is provided by the University of Edinburgh as part of the ÈIST project to generate transcripts and subtitle files from Scottish Gaelic audio and video recordings using automated speech recognition (ASR) technology.

Audio and video you submit will be used solely to generate the transcript and subtitle displays and the output files (JSON/SRT) you request. The University of Edinburgh does not use uploaded content for model training, profiling, marketing or any other secondary purpose.

The University uses Hugging Face to host this tool and carry out the transcription processing on its behalf. Hugging Face acts as a data processor for this activity, and processing takes place on infrastructure located within the EU.

Where you choose to use the optional English translation feature, the recognised text is additionally sent to OpenAI (ChatGPT) via EDINA's ELM platform for translation. The University of Edinburgh holds an enterprise agreement with OpenAI that includes a zero data retention arrangement: data sent for translation is not retained or used for training by OpenAI, and is discarded once the translation is returned.

We process this content in the University's legitimate interests in developing and providing Scottish Gaelic language technology, supporting an endangered language, and delivering the funded objectives of the ÈIST project (UK GDPR Article 6(1)(f)).

Please note that audio and video recordings may contain personal data, including identifiable voices. Before uploading any content, you are responsible for ensuring you have an appropriate lawful basis and any necessary permissions to process the content, including the consent or authority of anyone else whose voice or personal data it contains.

Sharing information

Other than the processors named above (Hugging Face and, for the optional translation feature, OpenAI via EDINA's ELM), your data is not shared with any third party.

This tool operates on a zero-retention basis. Audio, video and text are processed only transiently while your transcript or translation is generated and are not written to persistent storage. We do not retain uploaded content or generated transcripts or translations after your session ends. Data transmitted between your device and the tool is encrypted using HTTPS/TLS. Access to the application's backend, repository and administrative settings is restricted to authorised ÈIST project staff.

Transfers outside the United Kingdom

Transcription processing is carried out on Hugging Face infrastructure located within the EU. The EU is recognised by the UK as providing an adequate level of data protection, so no special transfer safeguard is required for this step.

If you use the optional translation feature, the recognised text will be transferred to OpenAI in the United States via EDINA's ELM platform. This transfer is covered by the UK–US Data Bridge and by Standard Contractual Clauses with a UK International Data Transfer Addendum included in OpenAI's Data Processing Agreement. Under the University's enterprise zero data retention agreement, OpenAI does not retain this data beyond the processing moment.

Automated decision making

This tool uses automated speech recognition and, optionally, machine translation to process your audio at your request. These are automated processes carried out to produce the output you have asked for. They do not involve profiling, and no automated decision is made about you that produces legal or similarly significant effects within the meaning of UK GDPR Article 22.

Further queries

If you have any questions about this privacy statement or the processing of personal data through this application, please contact the University of Edinburgh's Information Governance and Data Protection Office at dpo@ed.ac.uk.

Continued privacy notice

Continued privacy notice

← Back to the app