Copy link to clipboard
Copied
Hey Everyone,
I am reaching out to you with a question regarding data privacy concerns related to the use of the transcription tool in Adobe Premiere Pro. We utilize Adobe Premiere Pro in our university hospital to edit recorded (audio) interviews with educators. These interviews contain sensitive patient data as well as instructional information that must not, under any circumstances, be stored on external servers.
Therefore, we want to ensure that the transcriptions created with Adobe Premiere Pro are exclusively stored locally on our computers and not transmitted over the internet. Can you confirm for us that Adobe Premiere Pro's transcription tool does not send transcription data over the internet or store it on external servers unless explicitly enabled by the user?
We place great importance on safeguarding sensitive information and aim to ensure compliance with all data privacy regulations while utilizing your products. We appreciate your assistance and feedback on this matter.
Sincerely,
Marius
@Marius Ott A few years ago, we first introduced "Speech to Text" in Premiere Pro which could transcribe your sequence (not source clips). Your sequence audio was rendered to a temp file, uploaded, and transcribed on our servers, then the transcript was sent back down to Premiere Pro and stored in your project data.
In a release soon after, we switched to transcription on device with neither the audio nor the transcript text ever leaving your computer. I'm sorry I don't remember exactly when t
...@PolkaFever When transcribing in Premiere Pro, nothing leaves your computer and so therefore Adobe does not and cannot train on it. Also, the ML models doing the transcription on your computer are fixed and do not learn from or train on the audio you transcribe.
@Liam_Delap9563 Premiere Pro 2023, 2024, and 2025 do not use cloud processing for transcription. The earlier, first version of Speech to Text did, as I mentioned in my comment above. But in the versions 2023 and later, no audio or transcription data is sent to Adobe.
Copy link to clipboard
Copied
Copy link to clipboard
Copied
@Marius Ott A few years ago, we first introduced "Speech to Text" in Premiere Pro which could transcribe your sequence (not source clips). Your sequence audio was rendered to a temp file, uploaded, and transcribed on our servers, then the transcript was sent back down to Premiere Pro and stored in your project data.
In a release soon after, we switched to transcription on device with neither the audio nor the transcript text ever leaving your computer. I'm sorry I don't remember exactly when this switch happened but any version of Premiere Pro 2023 or 2024 works this way now. Now when you transcribe, the audio of your clips is rendered on device and passed to the transcription engine also on your device – nothing ever leaves your system.
The resulting transcript is stored in two places:
Copy link to clipboard
Copied
Thanks @mattchristensen for clarifying. Does this also mean Adobe does not use any data from transcripts for machine learning?
Copy link to clipboard
Copied
@PolkaFever When transcribing in Premiere Pro, nothing leaves your computer and so therefore Adobe does not and cannot train on it. Also, the ML models doing the transcription on your computer are fixed and do not learn from or train on the audio you transcribe.
Copy link to clipboard
Copied
Thanks @mattchristensen for your quick reply. Please update your speech-to-text FAQ about this and please elaborate on the use of server service in the FAQ's answer about GDPR regarding Speech-to-Text?
Copy link to clipboard
Copied
Adobe's Speech to Text feature in Premiere Pro does use cloud processing for transcription. The transcription data is sent to Adobe's servers unless you work completely offline. If data privacy is critical, consider using manual transcription or a fully offline tool. You may also want to check Adobe’s privacy settings and consult their support for confirmation.
Copy link to clipboard
Copied
@Liam_Delap9563 Premiere Pro 2023, 2024, and 2025 do not use cloud processing for transcription. The earlier, first version of Speech to Text did, as I mentioned in my comment above. But in the versions 2023 and later, no audio or transcription data is sent to Adobe.
Find more inspiration, events, and resources on the new Adobe Community
Explore Now