Home | News

10 Ways You’re Already Using AI

Evan Killham

Lifewire: Tech for Humans

Jun 8, 2023

social media icons on a phone - 1387373490

We're still pretty far from artificial intelligence on the level of the droids in Star Wars, but we've been living with various forms of it for years. Here are 10 examples of AI you might not have known you were using.

1. Browsing Social Media

You might have heard references to "the algorithm" on social-media platforms like Facebook, Twitter, and YouTube, and it usually has negative connotations. Changes to the algorithm—or its natural, mysterious behaviors – can affect traffic and visibility, which can, in turn, cause downturns in ad revenue and other monetization.

Social-media algorithms use information such as "likes," shares, subscriptions, content length, and other data to choose which posts or videos show up on other people's feeds. Companies typically keep it quiet how their algorithms work, partly to prevent people from gaming them, which can only make it more frustrating when they change.

Any app that has a section called "Recommended," "For You," "Posts We Think You'll Like," or something similar uses an algorithm that compares the activity on those posts to ones you've watched and interacted with to find things to put in that section. They may also let you accept or reject these recommendations, which helps the algorithm make better suggestions for you and other users.

2. Talking to Digital Assistants

If you've ever asked Siri to read a webpage for you, played a game with Alexa, or otherwise spoken to one of your electronics (and it responded), you've interacted with AI.

The digital assistants in your phone, tablet, smart speaker, computer, or TV are relatively simple compared to others, such as ChatGPT. However, they still work by analyzing voice data and deciding how to act. For example, when you tell Siri to turn on your bedroom lamp, it has to interpret that command to "know" what you're asking for and what to do.

03. Shopping Online

Just as social-media algorithms interpret user behavior to suggest posts, accounts, and videos, online stores use shoppers' behavior to make recommendations. They might look at items people often buy together and suggest them to people purchasing similar things.

Other information influencing these programs can also include shoppers' locations and time of day/year, You'll probably spot the results of this programming under a heading called "People Also Bought," "Recommendations Based on Your Purchase," or something similar.

4. Unlocking Your Phone

Your passcode isn't artificial intelligence, but if you use a biometric security method like a thumbprint or your face to unlock your phone, AI is behind it.

Apple's FaceID starts with you "teaching" your phone what you look like by showing it your face at different angles. Once it has that information, your phone can use it to ensure you are actually the person you're trying to unlock it. But it's not really a matter of comparing two pictures.

Yes, your phone has to compare the live data with the stored information. But it also has to account for different lighting, the angle you're holding the device and other factors. Has your beard grown? Are you wearing glasses or a mask today? Do you have different makeup on than when you set up your device? Biometrics like FaceID have to be able to pick out which information is relevant based on context, which is mainly possible through AI.

5. Driving, Walking, or Biking With an App

Google Maps, Apple Maps, Waze, and other navigation apps are also examples of artificial intelligence. When you ask them for directions, they typically offer the most efficient route based on current traffic conditions, construction, time of day, and other factors. Your path will also differ based on whether you're driving, walking, or biking, so you don't end up walking on a busy highway.

Many apps also include user-reporting functions. With these, travelers can report accidents and other unplanned obstacles to improve directions for other people.

6. Editing Photos

Over the past few years, Adobe has improved Photoshop's Object Selection and Quick Selection tools. These features let you pick out an object and separate it from the background with just a few clicks, and they're a huge improvement over older methods like meticulously painting layer masks.

However, you don't need professional software to take advantage of this AI-driven superpower. If you have an iPhone running iOS 16 or later, you can use the Photo Cutout feature in the Photos app to do the same thing. Tapping and holding on the subject of a photo creates an outline; from there, you can copy it for pasting somewhere else.

Both features come from "teaching" the processors how to distinguish between foreground and background based on studying thousands of examples.

7. Writing

Autocorrect and autocomplete features when you're writing an email or text message can occasionally be a liability when they malfunction, but they're impressive nonetheless.

Writing programs and apps study samples—especially the ones you provide them – to try to predict the word you mean to use. For example, over time, they might start to predict the word "you" should follow the phrase "I love," but depending on how you actually write, it might also predict "movies," a person's name, or "tacos."

8. Playing Video Games

Artificial intelligence has been a part of video games almost since their inception. Any time you play against a computer-controlled opponent, AI tells it what to do (although the competence of the programming can vary). The computer opponent in a chess video game bases its moves on having studied other games, and enemies in a shooter might react in different ways if you fire, run, or take cover.

One interesting, surprisingly complex example of early video game AI is in the arcade game Pac-Man, in which each ghost has different behaviors based on standing programming, the time you've spent playing, and, in some cases, their relationship to each other:

  • Blinky (red): Chases Pac-Man.
  • Pinky (pink): Aims for a few spaces ahead of Pac-Man.
  • Inky (blue): Target is relative to how close Blinky is to Pac-Man.
  • Clyde (orange): Chases Pac-Man until it gets within a certain radius and then runs away.
Pac-Man also had a "Scatter" mode that activated after a certain time. Each ghost heads to a different corner of the maze during this period. They alternate between "Chase" and "Scatter" modes as long as you play.

9. Listening to Music

Just like with shopping and social media, streaming music apps generate suggestions based on both what you and other people listen to. Each app calls its AI-driven curation something different, but it's usually something like a "station" or "radio."

You can often ask an app to generate a playlist based on a single song. This option typically looks at the activity surrounding that track, like what people who liked it also listen to, and populates a list in the same vein. It's similar to the "People Also Bought" area on a shopping site, but more rockin'.

10. 3D Printing

3D printing isn't as popular as streaming music or online shopping, but it does have a big user base. And every person in it uses AI.

When you prep a file to print on an FDM printer—one that extrudes plastic through a hot end and not one that uses resin—you'll feed it into a program that will "slice" it, meaning that it will create the individual layers that the printer will build up to make the final object. The slicing process takes the 3D model, breaks it into layers, and then creates commands that tell the printer where to move, how much plastic to extrude, where each layer should start, when and where to place supports, and more.

Each line of the code that eventually goes into your printer is based on the slicer's interpretation of the shape you fed into it. It's a more basic version of artificial intelligence, like the ghosts in Pac-Man, but it is nonetheless impressive.

FAQ

What is AGI in artificial intelligence?
AGI stands for "artificial general intelligence," and it's presently theoretical. While AI system can only perform tasks for which humans have programmed it, AGI (also known as "true AI") is entirely autonomous.

How does artificial intelligence learn?
Modern AI systems typically learn by example. Researchers feed huge amounts of data into them and then tell them what to "look" for. Some other AI systems, particularly in video games, are programmed with sets of instructions for different circumstances that they use to control characters, create procedurally generated levels, and other tasks.