IdeaScale AI security and admin settings
IdeaScale gives the admins and moderator control over the security of their Community Idea data while using IdeaScale AI to ensure the utmost data security and confidentiality. This can be accomplished by selecting the preferred AI model from the available options.
Admin Setting
Administrators and Moderators possess the capability to access the Admin settings, enabling them to choose between the Llama and OpenAI AI models for their analytics. It is important to note that this selection will influence only the admin or moderator's personal access to IdeaScale AI, without impacting others within the Workspace.
Llama: Llama is an open source AI model that runs on our servers, so your data never leaves the IdeaScale boundary. This is our most secure option. The default setting is configured to Llama.
OpenAI: OpenAl is our smartest Al option. Using this model will yield more dynamic and accurate results, but it means your idea data will leave the IdeaScale boundary temporarily.
Note: The OpenAI model is not enabled by default. To activate it, please reach out to your Customer Success Manager for assistance.
To access it, click on Admin option on bottom left corner of the page. A dropdown to select between the 2 models will appear.
Select the preferred option and proceed to confirm.
Once done, system will confirm the selection.
IdeaScale AI Security
IdeaScale AI provides admins/moderators with access to advanced AI capabilities directly within the IdeaScale platform. To ensure the utmost data security and confidentiality, the following measures have been implemented:
- Default Model: IdeaScale AI leverages the Llama 3.1 model as the default AI engine. Llama is an open-source model hosted securely within the IdeaScale data boundary. When utilizing the Llama model, user data remains exclusively within this secure environment at all times.
- OpenAI Integration: Users may optionally select the OpenAI model for their AI tasks. However, it is crucial to acknowledge that selecting OpenAI necessitates the temporary transmission of user data to the OpenAI API for processing. Upon completion, user data is promptly returned to the secure IdeaScale data boundary. This option is available upon consultation and approval from your designated account manager.
- Data Access and Privacy: IdeaScale AI does not proactively analyze or access community data. Data access is strictly limited to instances where a user initiates a specific AI-powered task, such as generating an analysis or report.
- Government Compliance: For government customers, the launch of IdeaScale AI will be strategically timed to ensure full compliance with all applicable FedRAMP requirements prior to deployment.
Please feel free to access the following links to learn more about the models we use: