TL;DR summary
If you only have a couple of minutes, here's what you really need to know.
Our current AI capabilities are:- Transcription
- Live Captions
- Quiz Me - a multiple choice question generator
- AI Outline - a summarization tool
✅ User data is used to deliver the service, improve the software, and maintain a secure environment.
✅ Approved third party providers can only use a user's data for specific purposes set by Glean in line with a Data Processing Agreement.
✅ The owner of content captured in Glean doesn't change: so if it's yours, it stays yours.
🚫 We don't sell or rent user data
🚫 We don't create our own AI models
🚫 We don't allow third parties to train their AI models on Glean data
🚫 We don't automatically record, transcribe, summarize, or generate quizzes for users.
Table of contents
- What Artificial Intelligence (AI) is there currently in Glean?
- Is Glean planning to add more AI capabilities in the future?
- How does Glean use the data captured in the software?
- How do third parties use data captured in the software?
- What is Glean's position on data ownership?
- How is data captured in Glean used to provide Transcription?
- How is data captured in Glean used to provide Live Captions?
- How is data captured in Glean used to provide Quiz Me?
- How is data captured in Glean used to provide AI Outline?
- Other useful information
What Artificial Intelligence (AI) is there currently in Glean?
When discussing AI capabilities in Glean, we mean:
- Transcription
- Live Captions
- Quiz Me
- AI Outline
To learn about how we handle data in relation to these specific features, please read on.
If you’re thinking about another feature or capability you have seen or heard about, please get in touch and discuss it with us at support@glean.co.
How does Glean use the data captured in our software?
We use data captured in the software to:
- provide the service to the user
- improve the software including carrying out quality assurance checks and repairing technical issues
- protect against security threats and illegal activity
- understand how users interact with our software
- improve user experience
To learn more, see our Privacy Notice here.
How do third parties use data captured in the software?
✅We use third parties to provide specific parts of our service, which means they need to handle personal data on our behalf.
✅ Third parties can only use the data for the specific purpose we allow and must always follow our instructions.
You can request the current list of third party processors we work with by emailing legal@glean.co.
✅We have a Data Processing Agreement in place with every third party that we work with that handles personal data.
🚫We don’t sell or rent data captured in the Glean software or, data derived from it.
🚫Glean doesn't create its own AI models
🚫 Our policy is not to allow third parties to train their own AI models on data captured in the Glean software.
What is Glean’s position on data ownership?
Glean follows the principle that ownership of data doesn't change regardless of it being captured within our software. So, if the content belongs to the user before it's captured in Glean, it remains theirs when it's in Glean.
For users who access Glean through their organisation or institution, it's their responsibility to make sure they follow the relevant local laws and their organisation's policies when capturing data within Glean. Learn more about this in Section 6 of the Glean End User Service Agreement.
How is data captured in Glean used to provide Transcription?
The Glean software has the functionality to generate a transcript of recorded audio. This option is available to users once recording has ended. We use a third party service provider called AssemblyAI to provide the transcription feature. AssemblyAI will only use the user's audio recording for the purpose of generating a transcript - and nothing more. Once the transcript has been provided to the user, the copy of the audio and the related transcript will be permanently deleted from AssemblyAI's systems.
🚫 We don't allow AssemblyAI (or any of our third party providers) to use Glean data to train their own AI models.
How is data captured in Glean used to provide Live Captions?
The Glean software also has the functionality to generate on-screen captions in real time. Glean uses a third party provider called AssemblyAI to provide the Live Captions feature. When an institution purchases Live Captions, the feature is then made available to users in their account.
If a user switches on live captions for an audio recording, a copy of the audio is sent to AssemblyAI who will then process the recording in order to generate the captions. Once the captions have been provided to the user, the copy of the audio recording and the generated captions are permanently deleted from AssemblyAI's systems.
🚫We don't allow AssemblyAI (or any of our third party providers) to use Glean data to train their own AI models.
How is data captured in Glean used to provide Quiz Me?
Glean's Quiz Me feature uses AI to generate multiple-choice questions based on a transcribed audio recording. To provide this capability, we use Microsoft Azure's Open AI service. This means we can benefit from Microsoft's industry leading experience in data privacy and security.
✅Microsoft hosts Open AI’s large language models on Microsoft Azure servers.
🚫The data that we give Microsoft in order to provide the Quiz me feature is NOT:
- sent to Open AI at any time
- used to train Open AI’s models
- used to improve Microsoft services
- used to improve any third party's product or services.
- used to improve the Azure Open AI model for other Glean users
Learn more about data, privacy, and security for Azure OpenAI Service here.
How is data captured in Glean used to provide AI Outline?
Glean's AI Outline feature uses AI to generate a simple summary based on a transcribed audio recording. To provide this capability, we use Microsoft Azure's Open AI service. This means we can benefit from Microsoft's industry leading experience in data privacy and security.
✅Microsoft hosts Open AI’s large language models on Microsoft Azure servers.
✅Transcripts are sent to Microsoft Azure servers and the AI model output is sent back to the user.
✅Microsoft store prompts and outputs for 30 days to identify patterns of service abuse and then deletes them.
🚫The data that we give Microsoft in order to provide the AI Outline feature is NOT:
- sent to Open AI at any time
- used to train Open AI’s models
- used to improve Microsoft/Azure services
- used to improve any third party's product or services
- used to improve the Azure Open AI model for other Glean users
Learn more about data, privacy, and security for Azure OpenAI Service here.
