Noelle RussellChief Executive Officer, AI Leadership Institute · 5x Microsoft Trustworthy AI MVP | 3.4+ million students | #1 Thought Leader in Agentic AI | AI-Powered Revenue Growth + Cost Optimization Expert ✨Best-Selling Author: Scaling Responsible AI | TEDx Sp

How AI Makes Our Jobs Easier

17 Sept 2024 · 2 min

Watch on YouTube

1:30

In this presentation, the speaker explains how applied AI technologies can streamline work tasks. They first illustrate computer vision using object detection with bounding boxes, showing how models can recognize items like chairs, people, and dogs at the pixel level. Building on that, they describe anomaly detection, using a parking‑lot scenario where a trained model expects an empty lot and flags a newly arrived car as an anomaly, triggering alerts for decision‑makers. The speaker then contrasts building custom models from scratch with leveraging commoditized applied AI services offered by major cloud providers (Amazon, Google, Microsoft). They argue that many foundational models—such as computer‑vision and natural‑language models—are now readily available through services like Microsoft Cognitive Services, allowing engineers to focus on integration rather than model development. The talk ends with a note that the speaker shared related resources on a Discord channel.

Chapters

computer vision. And you probably have seen these demos with like bound boxes, right? Where you look at something and the computer based on past data can look at something and go, oh, that looks like a chair and that looks like a person and that looks like a dog. So it's able to identify in computer vision, able to identify to the pixel level what something is. And then on top of that, you can add things like anomaly detection, right? So let's say you're looking at a parking lot. Right? And everything looks good. You've trained the model. This is what the parking lot should look like. It should be empty. And then a car rolls in. All of a sudden, an anomaly has occurred. There's a car in the parking lot. And so the, the decision maker, the, the model can now trigger an alert that says, I don't know if it's bad or good, but it's not an empty parking lot anymore. So there's so many things that you can do with this technology to just make your ability to do work faster. And so those are just a couple examples. There's dozens and dozens of, um, actually hundreds of models. You can build a model from scratch. So, which means that any one of us could create a new model. But today I'm a big believer in applied AI. Companies like Amazon, Google, Microsoft have all invested in building models for us to use as engineers. And those are called applied AI models. They're like computer vision, natural language. Like nobody's really building human language detection anymore. Like that's been done. And it's commoditized and available through services. I actually posted this in the Discord, um, like Cognitive Services for Microsoft.