Using AI But You Just Don't Know It

You are using AI

You are Using AI...You Might Not Know it!

It’s pretty simple to see Artificial Intelligence (AI) at work when you read a New York Times article about driverless car collisions or displaced factory workers. But, what’s harder is clocking the real ways that you are using AI in your day to day lives and just don't know it. Here are four everyday functions of AI technology that you may not notice until you take a closer look.

Advertising

Machine learning has revolutionized the advertising industry. According to a piece in Variety, Lexus teamed up with the AI powerhouse IBM Watson to watch over 15-years worth of “car and luxury brand campaigns that have won the Cannes Lions awards for creativity, as well as a range of other external data.” Then, they had the AI script an entire commercial with no human involvement. These kinds of script writing services aren’t just one off PR stunts, they are available for widespread commercial consumption.

The Marketing Artificial Intelligence Institute even noted that:

“Brands today are beginning to use commercially available artificial intelligence to intelligently identify and segment audiences, build ad creative, test variations, improve performance, and optimize spend—automatically, in real-time, and at scale.”

Your Email Inbox 

At the beginning of accessible email services (I’m talking Hotmail and MSN Messenger: The dark days), your junk mail was virtually indistinguishable from your regular mail. Today, ask anyone who uses Gmail about the last time they looked through their spam folder and see what happens. As is the case with most things that are hidden in plain sight, the spam folder in Gmail is real, but designed so that you aren’t very inclined to crack it open except to empty it out. While Gmail is blocking around 100 million extra spam messages a day. Google has implemented new “neural networks” into it’s spam filters that actually learn the language we use and can spot spam 99.9 percent of the time with only a 0.05 percent false positive rate (things that get labeled spam but aren't).

Using AI in Streaming/Music Curation 

There was once a very popular Spotify ad in New York City subway stations that read, “I want a partner who knows me as well as my Spotify Discover Weekly Playlist.” At one point in history, it may have been easy to think that some music lover somewhere would just comb through our most listened to albums and build a playlist for us. If that seemed too menial to be possible (which it 100 percent is) you could at least imagine that someone was at least going: “Well if they like this artist than they’ll like this genre” and letting a computer churn out the playlist for them. What you may not have imagined is that companies all over the world are working on AI that can actually listen to and interpret music. Not only so they can recommend it to you, but so that the AI will suggest other compositions you might like.

The application Brain.FM uses AI “to arrange musical compositions and add acoustic features that enable listeners to enter certain mental states.”  Amper Music does a similar thing, but analyzes other types of non-musical creative content in order to compose background or accompanying music. The company Asaii is even changing the inner record label landscape by seeking to eliminate the need for an A&R rep to scout talent by using machine learning algorithms. 

Online Translation Tools  

As of right now, punching in an English phrase and letting Google Translate churn it out in Spanish won’t get you a good grade, but thanks to Google’s Neural Machine Translation, that could soon change. The most basic form of any online translator basically takes a word in one language and uses a cache of words in another while trying to make a match.  But neural Machine Translation takes that to a whole new level. Not only is it learning how to translate word for word, but it is learning how to create more accurate translations based on the entirety of a sentence.  In other words, it is translating for context.

This could mean that while right now you may only be able to translate your English phrases into textbook Spanish sentences, in a few years you may be able to specifically translate your English phrases into any number of Spanish’s sub-dialects. As you read this Google’s Neural Machine Translation, rather than guessing based on statistics, is learning from a broad range of language inputs, linking them to grammatical constructs in language, and then deriving more relevant translations than a machine ever could in the past.