What are some interesting use cases of AI and machine learning? originally appeared on Quora, the place to gain and share knowledge, empowering people to learn from others and better understand the world. You can follow Quora on Twitter, Facebook, and Google Plus.

There are some very interesting use cases for artificial intelligence and machine learning in nearly every facet of our world. My field, marketing, is all about solving efficiency issues, improving intelligence gathering, and understanding so that we can predict and react faster than anyone else. Using NVIDIA (my employer) as an example, we have a lot of visual based User Generated Content (UGC) being shared across social media. Things like in-game screenshots from your favorite video game, architectural or product design simulations in Iray, and the list continues.

As you can imagine, the volume of content flowing in our direction can be massive. Especially in the video game world. And in terms of Community Management, a critical component to brand and social marketing, UGC identification, promotion, and nurturing is an essential task that is an ongoing requirement for a healthy community. Since there are thousands of images posted online in any given month by our fans, with peaks around game launches or contests, it is an almost entirely unfeasible ask of community managers to find, tag, store, and develop strategies with this amount of content.

In comes AI/Machine Learning.

Begin Scenario (hypothetical):

On May 2nd, 2017 , a fan named “Christian Johansen” posts on Instagram a picture of the computer he built with his own two hands. It was his first time building a computer, and he was extremely proud of his creation. He then spent a lot of time taking a picture. Click. “@NVIDIAGeForce #nofilter.” Post.

Using social media listening technology, companies like NVIDIA “listen” to the constant feed of internet postings looking for any time they are mentioned. We make long Boolean queries that say “Grab any mention of NVIDIA or its products on the internet, and pull it into a report so we can see what people are talking about.” Now, that’s an oversimplification, but that’s essentially what’s happening. That itself isn’t AI, but that’s just where this gets started.

Natural Language Processing for Sentiment Analysis:

As that “mention” gets pulled into the system, an AI called Natural Language Processing reads the post text and determines its “Sentiment.” Sentiment is used to determine if the post is positive, neutral, or negative, (and in some advanced cases, the emotion like “anger” “sadness” or “joy”). Doing this manually for every post that comes in isn’t feasible (we see tens of thousands of posts on any given week). AI does this automatically for us, and it can “learn” to improve its NLP Sentiment analysis as more posts pass through it, and as manual adjustments for errors are made. And speaking of Christian’s post, he used the “#nofilter” hashtag which can be assigned a “proud” tag since its basically saying “my picture was so good, I didn’t need to edit it.”

Image Processing, Computer Vision, Object Detection and Recognition:

This is where it gets cool. Through computer vision models, we can train a system to identify pictures that contain our products. This takes pictures pulled in from listening queries, passes them through the system to “see” what’s in it, and if it finds one of our GPU’s in it, for example, it will tag the image with the product name in our system.

Once that happens, we can set up a rule that says, “If an image has positive sentiment, ask the poster of that image if we can use it.” So, a message goes to the poster that says “Hey Christian! We loved your picture, is it cool if we share it and use it in some of our marketing materials? Reply with #NVIDIAYes or #NVIDIANo.” If he says yes, the system will download the image directly into our digital asset management system within our digital marketing platform, along with all the tags the AI applied and the post and author information as metadata.

Now, that’s just one use, but others can be along the lines of crisis management. Like, someone posts a picture of a car fire in the NVIDIA parking lot, the system can pull the image in, identify a fire on our property (based on post location data), and alert our onsite security to respond. So, this can help IDENTIFY, CLASSIFY, and NOTIFY.

If you want to see more examples of non-marketing focused use cases, NVIDIA has some listed here (not intended as spam, just literal examples): NVIDIA Deep Learning & AI Use Cases and Customer Success Stories.

This question originally appeared on Quora. More questions on Quora:

* Artificial Intelligence: Is A.I. an existential threat to humanity?

* Machine Learning: How much does machine learning matter in data science?

* NVIDIA: Why did NVIDIA win the GPU market?

Photo Credit: Paul Cooklin/Getty Images