Is AI taking over the world?
Is it secret? Is it safe?
Everywhere you look these days, someone is talking about artificial intelligence (AI). Whether you’re a die-hard Sci-Fi fan that believes AI will destroy the world or someone who thinks “Halleluiah! I can use it to summarize this long email thread”, you’ve undoubtedly been exposed to AI technology lately. People are using publicly available solutions that do everything from writing essays and blog posts to creating illustrations.
But what about AI for the corporate world? Using free AI-enhanced search engines like Bing or Bard is one thing, but you might not want your company data floating around the Internet. Microsoft has a solution called the Azure OpenAI Service that is designed especially for enterprise customers. Companies are already using it to expand translation capabilities or to improve customer service. But for those who don’t know about Azure OpenAI or have been hesitant to adopt the service, let’s define what Microsoft’s Azure OpenAI service even is.
The Azure OpenAI Service delivers advanced language AI with OpenAI GPT-3x and 4x, Codex, and DALL-E models with the security and enterprise promise of Azure. The Azure OpenAI Service is a result of a partnership between Microsoft and OpenAI. The service combines Azure's enterprise-grade capabilities with OpenAI's generative AI model capabilities.
That’s a lot of words to pretty much say - “It’s like OpenAI but for designed for enterprise customers”.
Azure OpenAI consists of four components:
Pre-trained generative AI models
Customization capabilities: the ability to fine-tune AI models with your own data
Built-in tools to detect and mitigate harmful use cases so users can implement AI responsibly
Enterprise-grade security with role-based access control (RBAC) and private networks
With Azure OpenAI, customers get the increased security capabilities of Microsoft Azure while running the same models as OpenAI. And all that data never leaves your Azure tenant. If you’re hesitant to adopt this new technology, we thought it would be great to share some of the concerns we’ve heard.
1. Is Azure OpenAI leaking my information out to the Internet?
Absolutely not. Your data is your data. The data is stored encrypted in your Azure subscription. The data is always under your control. Azure OpenAI, unlike commercial OpenAI, is not exposed to the open Internet. We also guarantee that your data is not used to train foundation models. Any fine-tuning you do stays in that Azure subscription, and never moves into the foundation AI models used by the world at large.
2. How can we be sure that Azure OpenAI is safe and secure? Can we monitor or track it?
The Azure OpenAI service is deployed within your Azure subscription, secured by you, accessed only by you, and tied to your datasets and applications. With Azure OpenAI Service, customers get the security capabilities of Microsoft Azure while running the same models as OpenAI. And like Azure itself, Azure OpenAI is SOC2, ISO, HIPPA, and CSA Star compliant. Your data is protected by comprehensive enterprise compliance and security controls available as part of Azure. You can take advantage of familiar foundational security components like:
Azure Identity and Access management (IAM) to give permission only to the people who need it.
Data encryption with customer managed keys
Control access with VNETs and private endpoints
Use Azure Monitor, Azure App Insights and more for metrics
Role-based Access Control (RBAC)
3. How can we prevent people from using Azure OpenAI with bad intentions?
It’s vital that AI is helpful and not harmful. OpenAI models are integrated with Microsoft-developed content filtering and abuse detection models. The content filtering system works by running both the prompt and completion through an ensemble of classification models aimed at detecting and preventing the output of harmful content. Customers can also use enterprise tools like DLP to monitor this type of activity.
We do recommend that each customer create their own AI governance Policy. The Microsoft Responsible AI Standard documentation can be a great starting point. We also have Microsoft’s Transparency Notes that are intended to help you understand how our AI technology works, the choices you can make that influence system performance and behavior, and the importance of thinking about the whole system, including the technology, the people, and the environment.
Here are the six Microsoft AI principles:
Fairness: Al systems shouldn't make decisions that discriminate against or support bias of a group or individual.
Reliability and Safety: Al systems should respond safely to new situations and potential manipulation.
Privacy and Security: Al systems should be secure and respect data privacy.
Inclusiveness: Al systems should empower everyone and engage people.
Accountability: People must be accountable for how Al systems operate.
Transparency: AI systems should have explanations so users can understand how they're built and used.
4. How do I know how much it will cost?
With AI being so popular these days, it’s normal for people to be concerned about the potential for runaway costs. Because the service is hosted in Azure, you can manage the cost like any other Azure workload. There is also a pricing calculator for Azure OpenAI on the Azure Pricing Calculator site.
As part of their partnership, Microsoft and OpenAI have agreed to work together on several key areas, including:
Developing AI technologies: The two companies will collaborate on the development of new AI models and algorithms, with a focus on natural language processing, computer vision, and other areas of AI.
Building an AI computing platform: Microsoft and OpenAI will work together to build a new AI computing platform that will allow developers to easily access and use advanced AI models and algorithms.
Advancing AI research: The two companies will also collaborate on a range of research projects aimed at advancing the state of the art in AI.
Making AI more accessible: The partnership aims to make AI more accessible to a wider range of developers and organizations, in order to bring the benefits of AI to more people.
Let us know if you have any questions, you’d like answered about Azure OpenAI. We’d love to hear from you.
Great post, thanks for sharing it