Edytuj

Udostępnij za pośrednictwem

Data, privacy, and security for Microsoft 365 Copilot in Viva Insights

  • Artykuł

Analysts can take advantage of the many features in Microsoft 365 Copilot for Viva Insights including suggestions for metrics, filters, and attributes to use in queries.

For example, Copilot can help analysts create queries by suggesting useful metrics that would be relevant to their queries.

How does it work on the backend?

Copilot in Viva Insights uses Large Language Model (LLM) technology, which allows users to perform language-based tasks by asking prompts in their natural language.

Microsoft uses these LLMs to match the user’s prompts to a set of metrics and "stories." A "story" is a product-defined group of calculations based on metrics, aggregations, and other variables.

What data does Copilot use to generate its responses?

Copilot in Viva Insights uses already-existing data within Viva Insights and the prompts the customer asks when they use the feature. It doesn't collect any new data.

Does Copilot in Viva Insights save or store any of the data it uses?

Copilot in Viva Insights doesn't log any of the user’s prompts. It saves only the metadata for each API call, such as the date and time of each prompt. The meta log can only be accessed by Viva Insights developers and restricted users.

How does Copilot in Viva Insights interact with Microsoft Graph and other Microsoft 365 services?

Copilot in Viva Insights doesn't interact in any way with any Microsoft 365 Copilot, Microsoft Graph, and other Microsoft 365 services.

How does Copilot handle data residency, data subject rights, and data deletion?

Copilot in Viva Insights follows the same data residency protocols currently in place for Microsoft Viva. For details, see the security policy for Microsoft Viva, and the Microsoft Viva compliance protocols.

How does Copilot in Viva Insights comply with GDPR and other privacy regulations?

Copilot in Viva Insights follows the same GDPR and privacy regulations as Viva Insights. For details, see the Viva Insights privacy guide for organization insights.

What safeguards does Copilot provide to address potentially offensive questions or responses or otherwise sensitive information?

Copilot in Viva Insights is designed with a robust filter system in place to proactively block offensive language and to prevent generating suggestions in sensitive contexts. Our commitment to continuous improvement drives us to enhance this filter system, making it smarter in detecting and removing offensive content generated by Copilot in Viva Insights. This includes addressing biased, discriminatory, or abusive outputs. We encourage users to report any offensive suggestions they come across while using Copilot in Viva Insights directly within the product. 

There are several high-risk use and sensitive cases that we aim to avoid: 

  • Sensitive use cases: Since the data & insights that Viva Insights can provide falls in a sensitive area, it’s difficult to objectively evaluate whether a particular analyst query question related to employee performance or workplace behavior patterns should be answered by Copilot or not. While we expect Copilot to answer a question like “How are employees in China collaborating?”, a query such as “Are employees in China working harder than employees in US?” might fall into the “Fairness” category. We're dedicated to addressing these complexities by blocking questions which can support making a “judgment”, while allowing questions that help users to obtain “facts.” 

  • Privacy Concerns: If the AI feature isn't adequately secured, it may pose a risk of exposing user data, including private trending themes and personal information, to unauthorized parties. Ensuring the highest level of privacy and security for our users is a top priority. 

  • Bias: The fairness and impartiality of AI systems like Copilot in Viva Insights depend on the quality and bias of the data they're trained on. If the training data contains biases, the AI feature could unintentionally generate content that reflects those biases, potentially causing harm or offense. We're dedicated to addressing bias in AI systems and working towards providing more equitable and inclusive outputs. 

By actively addressing these high-risk scenarios and working collaboratively with our user community, we're committed to delivering a safer, more responsible, and ethically sound AI experience with Copilot in Viva Insights.