AI: The future is now

There’s a big chance that you have read or heard the words Artificial Intelligence before, and when you do they may trigger emotions like fear or disgust (mostly due to movies... Terminator?), maybe curiosity or ultimately go like "whatever".

You could have heard it on the news, on social media, on a movie, a book or even in a coffee shop, literally it’s everywhere. I really meant everywhere. Let’s give you some examples. When you pay with your credit card it wouldn’t be a surprise if there’s some anomaly detection process to determine if it’s a fraudulent transaction. Here is another, you take a picture with the coffee you just bought and your phone recognizes faces on the picture and to who those faces belong, “does this face belong to my owner?” Asks to itself your device. “Yeah, let’s add it to this collection of pictures”. And another example, because you look awesome on the pic, you decide to post it on your social media page, and tada! “Here you go” says the social media “some ads of coffee shops”... They just recognized you are holding a coffee cup.

AI: Behind the scenes

Let’s now try explaining the technology behind the scenes. AI is divided into subtopics, one of them is Machine Learning, better known as ML. The best way to explain this is by a scenario or example. Programming in its traditional form consists of a bunch of nerds, sorry I meant to say a bunch of geeks, writing code to develop an algorithm that ultimately solves a problem. Now, this works out for a lot of problems but there are another ton of problems that this wouldn’t necessarily be the best way to go. Take for instance the task of recognizing an object in a photo. Coding an algorithm for that kind of task would take an extremely large amount of time and effort, even for the best programmers out there. Why? Well, code an algorithm to recognize a coffee cup… now what happens if the cup is upside down, will your algorithm detect it? What about adding the recognition of sunglasses? Because an application to only detect coffee cups isn't that useful. Luckily here is where ML can help. Think of it as another way of coding algorithms, where you have values for input and output (perhaps images of coffee cups and sunglasses as input and their respective label as output), after some analysis of the data you select a math equation to then tell the computer to train a model with the data at hand. That trained model is like an algorithm that was coded by a computer to solve a specific task (in our case detect coffee cups and sunglasses). That's why it’s a subtopic of AI, a computer just programmed for you, what more artificial intelligence than that?

AI: Available for everyone

But even though I tried to explain AI in a simple way, it still seems to be a really hard thing to do (at least for me). Some people could argue that there’s a lot of tools out there to make that process a little bit simpler, but still isn’t an easy task. Now, there's something I like to call AI integration. It basically consists of solving real problems faced by businesses with AI using services/applications already built by experts on the topic like AWS, GCP and Azure. Think of it as taking a shortcut on a very complex development path, where businesses benefit from AI without all the overhead of building one from scratch. Saying it’s a plug and play solution would be a lie, but it would feel like it if you have tried the other way around before. If you are planning on using AI for a project or your business, I recommend you to take a look at what services already exist, you may find something that works. Or better yet contact us, we would be more than happy to help.