Privacy

Access requires an ASU CreateAI account, so the service is limited to ASU faculty, staff, and approved university projects. It is not offered to the public and is not intended for use outside the university.

Only what signing you in and running your jobs requires. The tool keeps no profile of settings or defaults for you: choices such as the page title and stylesheet apply to the document you are generating and are not remembered between jobs.

• ·Your ASU identifier (asurite) from CreateAI sign-in, which is what your work is filed under, plus your display name and the time you last signed in.
• ·A sign-in cookie that your browser sends back to us. It is httpOnly and SameSite=Lax, and it is marked Secure in production, so page scripts cannot read it.

The media file you upload is stored so the job can run and so a repeat upload of the same file can reuse the earlier result instead of paying to transcribe it twice. Alongside it we keep the file name, type, size, and a SHA-256 hash of the contents.

Everything the tool produces from that file is stored too:

• ·The transcript, both the raw model output and the cleaned-up version.
• ·The generated accessible HTML page, including the page title and stylesheet URL you chose for it, which are written into the file itself.
• ·Any translations you request, in up to 8 languages.
• ·A write-once snapshot of the original AI output per language, kept so an edited transcript can still be compared against what the model first produced.
• ·Job metadata: status, timestamps, how long processing took, the AI-generated title, and any error code.

Usage is counted per month and in total: number of jobs, minutes of audio, cost, and input and output tokens.

When you edit a transcript we record how much changed against that original AI snapshot: word counts added, removed and changed, the percentage changed, a similarity score, which model produced the text, and the optional one-click category you pick for what you were fixing. This is how we tell whether the AI is doing well on real material. It describes the edit, not you, and we do not use it to evaluate people.

Counts of translation runs and edits are also published as CloudWatch metrics, as numbers only, with no content or identifiers attached.

Transcription, translation, and transcript cleanup run on the ASU CreateAI platform, using your own CreateAI credentials. Your audio and text are sent there to be processed, and CreateAI's own terms and privacy practices apply to that step.

Files and data are held in ASU-controlled AWS infrastructure. We do not sell your data, we do not use it for advertising, and we do not share it with anyone outside the university for their own purposes.

Only you. Every request checks that the job belongs to the account asking for it, and a job with no recorded owner is denied rather than shared. Administrators can reach operational data such as usage counters and error logs for support and cost management.

Deleting a job removes it for real: the in-flight record, the stored job entry, and the stored files, meaning the uploaded media, the transcripts, the HTML, and the translations. The deduplication pointer is cleared at the same time, so re-uploading the same file starts a genuinely new job.

Anything you have not deleted is kept for as long as the service runs, since there is no automatic expiry today. If you want your whole history removed, ask the team.

Besides the sign-in cookie, the app keeps a few small values in your browser: the job you currently have open, the date you last looked at the What's New page, and an optional flag that turns on extracting audio from video before uploading. Clearing your browser storage removes them.

This page describes how the software behaves, and it is not a legal notice from the university. Arizona State University's own privacy and data-handling policies apply here in full and take precedence over anything written here.

For questions about this page, or to request removal of your data, contact the Learning Enterprise team that runs the service.