Microsoft has unveiled its new Trusted AI narrative, a key step in solidifying the message around the commitment and capabilities that ensure safety, privacy and security across the Microsoft cloud.
These updates underscore a focus on creating robust solutions that protect users and data within the Azure artificial intelligence (AI) ecosystem. Here are the key announcements related to Trusted AI in Azure.
News on trusted AI in the field of Azure Security
Here are the key updates that reinforce Microsoft’s commitment to AI security.
-
Risk and security assessment for indirect prompt injection attacks
This new functionality allows users to simulate prompt injection attacks in their generative AI applications and measure the failure rate of detecting and mitigating these attacks. Users can drill down into the details of the assessment to better understand how their application responds and the associated risks. Learn more here.
-
Unsafe content detection and remediation in Azure AI Content Safety
This new advanced capability not only detects unsubstantiated content or hallucinations in AI-generated outputs, but corrects them in real time. By aligning generative responses with connected data sources, the correction ensures that the final results are accurate, reliable and soundly based. Learn more here.
-
Protected material detection for development code
Now available in Azure AI Content Safety, this feature helps identify and block the use of pre-existing code by checking for matches in public GitHub repositories. This tool encourages transparency and collaboration within the development community. Read more here.
-
Public security preview of embedded content
This feature provides a robust solution for developers who want to embed generative AI capabilities directly into their devices. Read more here.
-
New quality assessment capabilities in Azure AI Studio
These new capabilities include evaluations for protected material in outputs and new math-based quality metrics. In addition, they allow you to simulate typical user interactions with the application, thus generating high quality tests. Learn more here.
What’s New in Azure Trusted AI Privacy
In this section we will look at major updates that reinforce Microsoft’s commitment to privacy in Generative AI.
-
Encrypted information with the Whisper model in Azure OpenAI Service
This capability is designed for developers in highly regulated enterprises and industries, allowing them to process sensitive voice transcription data in the cloud, keeping the information encrypted even during processing.
This is crucial for industries that require high levels of security in managing sensitive data. Read more here.
-
Overall Availability of Confidential Virtual Machines (VMs) with NVIDIA H100 Tensor Core GPUs
With these VMs, Azure enables confidential computing on GPUs, giving customers verifiable assurance that their data and code is protected throughout its lifecycle, even while processing in CPU or GPU memory.
This opens up new possibilities for privacy-preserving multiparty analysis, model training and collaboration with third parties. Learn more here.
-
Azure OpenAI Data Zones for EU and USA
Coming soon, this functionality will simplify the management of generative AI applications, enabling regional data processing while ensuring data residency within specific geographic boundaries.
This will give customers greater access to models with more efficient data processing. Find out more here.
The future of trusted AI
With these updates, Microsoft is committed to the development of secure, private and trusted AI for businesses of all sizes. By doing so, we ensure that generative AI applications on Azure meet the highest standards of data protection and security, providing peace of mind to organizations adopting these innovative technologies.
At Bravent, we understand the importance of implementing reliable AI solutions that meet the specific needs of each business. Our team of experts is ready to help you integrate these new capabilities into your AI projects and ensure that your company is always at the forefront of technological innovation.
Contact us today!
Ready to take your AI project to the next level? At Bravent, we’re here to offer you the support and expertise you need.
Contact us now to find out how we can help you implement reliable and secure AI solutions in your organization.
For more details, you can contact us at Info@bravent.net
