Organizations throughout industries are leveraging Microsoft Azure OpenAI Service and Copilot companies and capabilities to drive development, improve productiveness, and create value-added experiences. From advancing medical breakthroughs to streamlining manufacturing operations, our clients belief that their information is protected by strong privateness protections and information governance practices. As our clients proceed to develop their use of our AI options, they are often assured that their beneficial information is safeguarded by industry-leading information governance and privateness practices in essentially the most trusted cloud in the marketplace at the moment.
At Microsoft, we now have a long-standing observe of defending our clients’ info. Our strategy to Accountable AI is constructed on a basis of privateness, and we stay devoted to upholding core values of privateness, safety, and security in all our generative AI merchandise and options.
Microsoft’s current privateness commitments lengthen to our AI industrial merchandise
Business and public sector clients can relaxation assured that the privateness commitments they’ve lengthy relied on for our enterprise cloud merchandise additionally apply to our enterprise generative AI options, together with Azure OpenAI Service and our Copilots.
- You’re in charge of your group’s information. Your information shouldn’t be utilized in undisclosed methods or with out your permission. Chances are you’ll select to customise your use of Azure OpenAI Service by opting to make use of your information to superb tune fashions on your group’s personal use. When you do use your group’s information to superb tune, any fine-tuned AI options created along with your information shall be obtainable solely to you.
- Your entry management and enterprise insurance policies are maintained. To guard privateness inside your group when utilizing enterprise merchandise with generative AI capabilities, your current permissions and entry controls will proceed to use to make sure that your group’s information is displayed solely to these customers to whom you might have given acceptable permissions.
- Your group’s information shouldn’t be shared. Microsoft doesn’t share your information with third events with out your permission. Your information, together with the information generated via your group’s use of Azure OpenAI Service or Copilots – equivalent to prompts and responses – are stored personal and will not be disclosed to 3rd events.
- Your group’s information privateness and safety are protected by design. Safety and privateness are included via all phases of design and implementation of Azure OpenAI Service and Copilots. As with all our merchandise, we offer a powerful privateness and safety baseline and make obtainable further protections that you would be able to select to allow. As exterior threats evolve, we are going to proceed to advance our options and choices to make sure world-class privateness and safety in Azure OpenAI Service and Copilots, and we are going to proceed to be clear about our strategy.
- Your group’s information shouldn’t be used to coach basis fashions. Microsoft’s generative AI options, together with Azure OpenAI Service and Copilot companies and capabilities, don’t use your group’s information to coach basis fashions with out your permission. Your information shouldn’t be obtainable to OpenAI or used to coach OpenAI fashions.
- Our merchandise and options proceed to adjust to world information safety rules. The Microsoft AI merchandise and options you deploy proceed to be compliant with at the moment’s world information safety and privateness rules. As we proceed to navigate the way forward for AI collectively, together with the implementation of the EU AI Act and different legal guidelines globally, organizations could be sure that Microsoft shall be clear about our privateness, security, and safety practices. We are going to adjust to legal guidelines globally that govern AI, and again up our guarantees with clear contractual commitments.
You will discover further particulars about how Microsoft’s privateness commitments apply to Azure OpenAI and Copilots right here.
We offer packages, transparency documentation, and instruments to help your AI deployment
To help our clients and empower their use of AI, Microsoft affords a variety of options, tooling, and assets to help of their AI deployment, from complete transparency documentation to a collection of instruments for information governance, threat, and compliance. Devoted packages equivalent to our industry-leading AI Assurance program and Buyer Copyright Dedication additional broaden the help we provide industrial clients in addressing their wants.
Microsoft’s AI Assurance Program helps clients be certain that the AI purposes they deploy on our platforms meet the authorized and regulatory necessities for accountable AI. This system contains help for regulatory engagement and advocacy, threat framework implementation and the creation of a buyer council.
For many years we’ve defended our clients towards mental property claims regarding our merchandise. Constructing on our earlier AI buyer commitments, Microsoft introduced our Buyer Copyright Dedication, which extends our mental property indemnity help to each our industrial Copilot companies and our Azure OpenAI Service. Now, if a 3rd get together sues a industrial buyer for copyright infringement for utilizing Microsoft’s Copilots or Azure OpenAI Service, or for the output they generate, we are going to defend the shopper and pay the quantity of any adversarial judgments or settlements that consequence from the lawsuit, so long as the shopper has used the guardrails and content material filters we now have constructed into our merchandise.
Our complete transparency documentation about Azure OpenAI Service and Copilot and the shopper instruments we offer assist organizations perceive how our AI merchandise work and present decisions our clients can use to affect system efficiency and habits.
Azure’s enterprise-grade protections present a powerful basis upon which clients can construct their information privateness, safety, and compliance methods to confidently scale AI whereas managing threat and guaranteeing compliance. With a variety of options within the Microsoft Purview household of merchandise, organizations can additional uncover, defend, and govern their information when utilizing Copilot for Microsoft 365 inside their organizations.
With Microsoft Purview, clients can uncover dangers related to information and customers, equivalent to which prompts embrace delicate information. They’ll defend that delicate information with sensitivity labels and classifications, which suggests Copilot will solely summarize content material for customers once they have the best permissions to the content material. And when delicate information is included in a Copilot immediate, the Copilot generated output robotically inherits the label from the reference file. Equally, if a consumer asks Copilot to create new content material based mostly on a labeled doc, the Copilot generated output robotically inherits the sensitivity label together with all its safety, like information loss prevention insurance policies.
Copilot dialog inherits sensitivity label
Lastly, our clients can govern their Copilot utilization to adjust to regulatory and code of conduct insurance policies via audit logging, eDiscovery, information lifecycle administration, and machine-learning based mostly detection of coverage violations.
As we proceed to innovate and supply new sorts of AI options, Microsoft will proceed to supply industry-leading instruments, transparency assets, and help for our clients of their AI journey, and stay steadfast in defending our clients’ information.
