Group Transcribe FAQ

Product

What are you launching and what are its features?
Group Transcribe is a free app on iOS that provides real-time speech-to-text transcription and translation. This app was designed for use by users joining sessions from their own phones while they are in the same room. Group Transcribe has higher accuracy and speaker attribution for in-person meetings based upon the volume of the speaker captured by the microphone of each phone used in the meeting. Users can see who said what in real-time from their phones, in their own language.
Why should I use the Group Transcribe app?
The Group Transcribe app is intended for people who would like to use transcription or translation during conversations or meetings and people who are not native speakers. Note-taking pulls your attention away from the speaker and makes it hard for you to follow the conversation. If a meeting or conversation has non-native speakers, it is challenging for them to participate in the conversation because of the language barriers. Expensive conference room hardware requires specific setup and is not mobile. We see a need for a simple, free, and yet quality solution for in-person conversation transcription and translation.
- Real-time transcription and translation accuracy by speaking into your own device
  - Accurately captures what is being said by speaking into your own device using your own device’s microphone while everyone is in the same room
  - Text appears on your phone showing who said what in real time
- Lets you be more productive at work and study
  - Save time when summarizing in-person conversations, meetings, interviews, work, or tutoring sessions and project discussions
  - Have all your past transcripts at your fingertips. Transcripts of your meetings are stored on your device so you can easily access and re-review your previous conversations, meetings, interviews, work, or tutoring sessions and project discussions
- Lets you be more present in a conversation
  - Follow the speaker, understand discussions, and fully participate in conversations
  - Speak your language naturally and freely without pushing-to-talk
  - Read the transcript in your language of choice
Can I see and export or share previous transcripts?
Yes, you can see all transcripts for meetings that you were a participant or host on your home screen. You can export and share the transcripts after the session is over.
Does each person in the conversation need to use their own phone?
Yes, Group Transcribe requires each person in the conversation to be in the same room and to use their own Group Transcribe app for transcription or translation. If there are participants to a meeting who do not have their own app, this will impact the transcription attribution accuracy and those participants speech will be attributed to other speakers.
How does the Group Transcribe app work?
A transcription session starts when one person, a.k.a. a meeting owner, initiates a meeting from their Group Transcribe app. They can invite other people in the same room to join by Bluetooth, QR code, or links. Once other participants receive the invitation, they can use it to join the same transcription session from their Group Transcribe app. As each person speaks to their own phone, they can see their own transcript real-time. Each sentence is marked clearly by the display name of speaker so you can follow the conversation fully.
If my friend and I are not in the same room, can we still use Group Transcribe?
While it is possible, this is not the intended use for this App. The transcription and translation quality such as speaker attribution will be discounted because the app is designed for people to use the app in the same room.
Can I use Group Transcribe for online meetings?
While it is possible, this is not the intended use for this app. The transcription and translation quality, such as speaker attribution, will be discounted because the app is designed for people to use in the same room.
How can I join a meeting or conversation on Group Transcribe?
You can join by Bluetooth, QR code, or links.
What is a nearby meeting and how can I find nearby meetings?
The nearby meetings feature allows hosts of a meeting to let others join your meeting by clicking the nearby meeting ID on the start page. You need to enable Bluetooth for the meeting. Please keep in mind that if you enable this function, anyone with internet access nearby may be able to join your meeting.
Can I turn off the nearby meeting function by default, or when I am in a conversation?
You can turn off the nearby function by turning off the Bluetooth access of the app by default. If you are in a conversation and don’t want others to find the meetings you are in, you can turn off the nearby meeting function by clicking the icon next to the meeting ID to disable nearby meeting function.
Is there a limit on the number of people who can join by Bluetooth?
We have set a limit of four people to join by Bluetooth. After that, the meeting owner needs to manually turn the Bluetooth back on to enable others to join by Bluetooth.
Can I see who is in the meeting at any time?
You can see the real-time number of devices as well as the display names of the devices in a meeting from the home screen.
Is there a limit to the number of people who can join the meeting?
While there is no limit, Group Transcribe works best for in-person meetings up to four people.

Privacy

Do I need to have an account to use Group Transcribe?
No. Group Transcribe doesn’t require you to be authenticated (e.g. with a user ID and password).
How does Group Transcribe attribute speakers during meetings?
Group Transcribe is designed for in-person meetings where all the devices are in close proximity to each other. By comparing the level of a person’s voice volume on each microphone across devices, the cloud service attempts to identify which device is closest to the speaker and the language preferences associated with that speaker. Each app in use will send data to the service and record all speakers. Group Transcribe does not create or use speaker recognition identifiers of meeting participants.
How does Group Transcribe use my meeting data?
Group Transcribe sends audio and text input from each app in use to Microsoft’s online speech recognition and translation technologies to transcribe, attribute, and translate multi-speaker conversations. Your display name is stored on your device only. When you engage in a meeting session, a randomly generated identifier associated with your display name and device is generated for the session and used for the purposes of attributing your speech and returning the transcript to each app. The resulting transcripts, customized for each participant’s chosen translation setting for the session, are shared with each meeting participant. Microsoft’s cloud services will process your meeting recordings and related transcripts for the duration of the meeting session. By default, we do not retain your meeting recordings and related text transcripts. Transcripts of meetings you have participated in are stored on your device. Meeting participants can opt in to contribute their meeting sessions to help Microsoft improve its speech recognition technologies.
If I agree to contribute my meeting recordings, what data do you collect and how do you use it?
If you decide to contribute meeting recordings, Microsoft will retain the audio and related speech recognition generated text transcription of your meeting only if if all participants within a given session have similarly agreed to contribute their meeting recordings. The audio and related speech recognition generated transcriptions of your meeting recordings will help us understand multi-speaker conversations to make Microsoft’s online speech technologies, such as speech recognition and speaker attribution during conversations, more accurate for you and everyone who speaks your language. If you decide to contribute meeting recordings, some of these recordings will be reviewed by Microsoft employees or contractors from other companies who work for Microsoft to improve Microsoft’s speech recognition technologies. Examples of how your meeting recordings could improve Microsoft speech technology, include the following:
- Recognize words when multiple speakers are talking at the same time and provide more accurate speaker attribution.
- Understand how language changes when people collaborate.
- Recognize speech from different speakers using a variety of devices despite background noises.
- Recognize complex and nuanced aspects of how people talk – like accents or regional dialects, and how sentences are structured in different languages.
Who has access to contributed meeting recordings and how does Microsoft protect my privacy?
Only permitted Microsoft employees and vendors working on behalf of Microsoft who have agreed to non-disclosure agreements and are required to meet high privacy standards will be able to review your voice clips in order to build and improve our speech recognition technologies. Microsoft takes steps to de-identify meeting recordings. For example, Microsoft does not collect and retain any account credentials or other persistent device IDs— so meeting recordings aren’t linked directly to you or other meeting participants. Display names that were associated with transcripts are not retained. Our automated software also takes steps to de-identify the meeting recordings by removing long strings of numbers that could be phone numbers or credit card numbers and anything that looks like an email address. In addition to the automated de-identification steps above, reviewers do not have access to the full meeting sessions and only listen to smaller, randomized snippets of audio recorded during different donated meetings so that the full content and context of the conversation is obscured to these reviewers. And if a reviewer discovers personal information that could be used to identify you, we remove that portion of the meeting from the recording and transcript used for product improvement. Meeting recordings contributed to Microsoft are only retained for product improvement if all participants within a session have agreed to contribute. Please be aware that that if you allow people other than yourself to participate in your Group Transcribe meeting session that you have agreed to contribute, you’re responsible for ensuring that they understand and consent to Microsoft using the meeting recordings as described here. Also note that if other sounds are audible when a meeting session is recorded, those sounds might be recorded as well. You can change your mind at any time in Settings or delete individual meeting recording donations in Settings > Manage Meeting Recording Donations.
What controls do I have over meeting conversations donated?
You can delete individual meeting recording donations anytime you want within your app under Settings > Manage Speech Donations. The list shows all the past meeting sessions you agreed to donate. You can select and delete any of them at any time. This will delete the audio and transcription meeting recording from Microsoft’s cloud for the relevant session. Meeting transcripts stored on your device in >Setting>History will not be deleted.
Where does Microsoft store meeting recordings contributed for product improvement and how long will they be retained?
Meeting recordings are stored in secure, encrypted servers. Microsoft retains meeting recordings that have been contributed to be sampled and reviewed for up to two years, during which time they may be sampled for product improvement. If we do sample, review, and transcribe the meeting recording, it may be kept for longer than two years in order to train and improve our speech recognition technologies.

To learn more about Microsoft and your privacy, please see the information in the Microsoft Privacy Statement https://go.microsoft.com/fwlink/?LinkId=521839