To many people, artificial intelligence (AI) seems like a futuristic — and possibly frightening — concept. But the truth is that many Web services and technology companies have been quietly building AI capabilities into their products for some time. In fact, if you've been online or used a mobile device today, you probably encountered an AI, perhaps without realizing it.
People use the term artificial intelligence to describe a wide range of different capabilities. In general, however, it refers to computers that think and learn in ways similar to humans. As one example of AI, many of today's AI systems have natural language processing capabilities, which means that they can understand spoken or typewritten words the way humans typically use them. Instead of requiring inputs in a particular format, they can answer ordinary questions, like "What will the weather be like today?" or "What is the capital of Venezuela?"
Other AI examples that have become more common include voice recognition (the ability to turn verbal communication into written text), image recognition (the ability to identify what is in a picture) and machine learning (the ability to learn without being explicating programmed). And many AI systems are closely related to big data analytics, particularly predictive analytics.
So where might you encounter these AIs? Here are twenty examples of artificial intelligence that people — probably including you — are using every day.
According to Amazon, its Echo and Echo Dot were some of this year's best-selling holiday gifts. In fact, the company says it sold nine times as many of the IoT devices this year as last year.
If you were one of the many who purchased these speakers with voice recognition capabilities, you're familiar with Alexa, the AI that powers the devices. It also runs Fire TV devices, and Amazon also offers an SDK that allows developers to build Alexa into their own devices. In addition, Amazon started the Alexa Fund to invest $100 million in startups that are incorporating the technology into their products.
Because it is built into Windows 10, Cortana may be the most widely accessible AI available today. Named after a character from the Halo video games, the personal assistant is also integrated into the Xbox One game console, Skype, Microsoft Band and the Edge browser, and Microsoft has also released versions for Android and iOS.
This interactive assistant responds to simple questions and voice commands. It excels at tasks like setting reminders, tracking packages, opening apps, sending emails and texts, managing your calendar and performing Web searches for specific facts. It also includes a music recognition engine.
When it debuted on the iPhone 4s in October 2011, Siri was the first personal assistant software with voice recognition capabilities that most people had ever used. While it was far ahead of anything available at the time, the early versions had some shortcomings, including frequently misunderstanding voice commands. In July 2014, Apple fixed many of those problems by quietly migrating Siri to a neural net platform. The smart assistant now has machine learning capabilities that allow it to get better at handling user requests over time.
In 2015, Apple revealed that Siri gets more than 1 billion requests per week.
Like Microsoft and Apple, Google also has a personal assistant with AI capabilities, but it isn't as widely available as Siri or Cortana. Announced last year, Google Assistant is currently only accessible to people who own a Pixel smartphone or a Google Home speaker (which is similar to Amazon's Echo) or who use the Allo instant messaging app.
First announced in the spring of 2016, Google Assistant grew out of the Google Now service (see below) that was integrated into Android. Like other smart assistants, it can answer simple questions, play music, manage schedules, make dinner reservations, get directions and more. It also integrates with Web services like Pandora, Uber, Spotify, Netflix, YouTube and others.
5. Google Now
According to NetMarketShare, 65.87 percent of all mobile devices in the world run the Android operating system. Nearly all of them can run Google Now, the predecessor to Google Assistant. However, unlike most other digital assistants, while Google Now accepts voice commands, it doesn't respond aloud. Instead, it replies with "cards" that provide the user with the requested information.
Interestingly, Google Now's AI capabilities also allow it to learn from users' previous activities and predict what information they might want next. For example, if you read a news article about a particular sports team, it will push out notifications regarding that team's latest game, or if you regularly drive to a particular place at a particular time, it will provide unsolicited traffic information for those destinations.
Google seems to be phasing out the "Google Now" designation, and it's possible that Google Assistant will completely replace this earlier AI.
If you've conducted a Google search today, you probably interacted with an AI. Google is famously secretive about the inner workings of its Web search service, but in 2015, it revealed than an artificial intelligence called RankBrain helps determine which websites appear in response to searches. At that time, the company said that a “very large fraction” of its searches were handled by RankBrain, and it's reasonable to assume that the company has come to rely even more heavily on its AI since then. The tool allows the search engine to learn from unique queries, and the company has said that it has become one of the most important determiners of a website's ranking on the search page.
Last year, Facebook CEO Mark Zuckerberg made — and kept — a New Year's resolution to build an AI for his home. But that was far from the only AI project Facebook personnel were involved in. The company is heavily invested in AI research, and it uses artificial intelligence to determine what shows up in your Newsfeed, to recognize faces in photos and videos, and to generate automatic captions for video content. The social network also recently said that it would use AI to help weed out the fake news circulating on its service after untrue stories became a big issue in the 2016 elections.
In a blog post, Twitter CEO and co-founder Jack Dorsey wrote, "Machine learning is increasingly at the core of everything we build at Twitter. It’s powering much of the work we’re doing to make it easier to create, share, and discover the very best content so that every time you open Twitter you’re immersed in the most relevant news, stories, and events for you." It has a research arm called Twitter Cortex that focuses on artificial intelligence technology. The company has also purchased several AI startups in recent years in order to expand its capabilities.
Like Facebook and Twitter, LinkedIn has also been heavily involved in AI research and has acquired some smaller machine learning firms. It does have one AI service that sets it apart from the other social networks, however. If you have ever looked for a job or for a job candidate on LinkedIn, you have used an AI. Matching job seekers with firms that are hiring has become one of LinkedIn's most lucrative services, and AI gives it predictive capabilities that the company hopes will set it apart in the very competitive job search market.
Recommendation engines have become such a ubiquitous part of the online world that most of us take for granted that services like Netflix will suggest options for us based on our past history. What might be more surprising, however, is how deeply these AI engines are integrated into the monetary side of the business. In a paper, Netflix executives said that the service's AI recommendation engines save the company $1 billion per year by helping it dramatically reduce its churn rate and keep people binging for hours and hours. In addition, a separate article revealed that the company uses an AI to help determine how much it should be paying for new content. If you've ever wondered how Netflix can afford to make all of the original content that it has been churning out recently, perhaps you should ask the AI.