From the course: Microsoft Azure AI Fundamentals (AI-900) Cert Prep by Microsoft Press

Unlock this course with a free trial

Join today to access over 24,900 courses taught by industry experts.

Describe considerations for reliability and safety in an AI solution

Describe considerations for reliability and safety in an AI solution - Azure AI Services Tutorial

From the course: Microsoft Azure AI Fundamentals (AI-900) Cert Prep by Microsoft Press

Start my 1-month free trial Buy for my team

Describe considerations for reliability and safety in an AI solution

“

- [Instructor] Where in the responsible AI and how do we ensure that AI is reliable and safe? That the software performs consistently and safely. Well, we test. Measure twice, cut once. Well, never stop measuring. (chuckles) Test under real-world conditions, for sure. And here, when we think of stress-testing the AI, I think of load balancing and the fact that Microsoft Azure has effectively limitless compute. That's one of the nice things. It's tough to really stress test these APIs 'cause they're built for speed. But nonetheless, you want to think about redundancy and fallback mechanisms. If your AI misfires or has a failure to launch, how do you escalate to a human quickly? You see? Therefore, we want to make sure that in our CI/CD, continuous integration and continuous delivery processes, we're monitoring AI in production just as much as we are any other component, looking for anomalies and looking to retrain models if needed. Chances are, yes, we will periodically need to do…

Contents