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.
Like Netflix, music services are also heavily invested in making good recommendations. Apple famously has human-curated recommendation lists, but Spotify is betting on AI. Its Discover Weekly feature makes custom recommendations for each user based on his or her past listening history, and those recommendations use machine learning to get better with time. Of course, the service also allows users to create custom playlists and share them with friends, but the AI-based feature is an attempt to introduce users with music they haven't heard before that they may like.
A quick search turns up the fact that Spotify has several job openings related to machine learning and other aspects of AI, so it's safe to assume the company is continuing to pursue its AI investments.
Like other streaming services, Pandora also uses algorithms to power its recommendation engines, but it doesn't discuss its artificial intelligence and machine learning capabilities. Its Thumbprint Radio feature, which it describes as a hyper-personalized radio station, does seem to have some machine learning capabilities, however, as it tracks songs that users have given a "thumbs up" and uses them to find similar songs. The service also has a Music Genome Project which uses a wealth of data to find similarities between different pieces of music.
You've probably never heard of DSSTNE (pronounced like "destiny"), but you have almost certainly used it. DSSTNE is Amazon's recommendation engine for its online shopping site; in other words, it's the service that tries (often successfully) to cross-sell products on Amazon.com. Last year, Amazon released DSSTNE under an open source license, which means that the same technology may now be integrated into other services, applications and websites that you use regularly. The company said, "We are releasing DSSTNE as open source software so that the promise of deep learning can extend beyond speech and language understanding and object recognition to other areas such as search and recommendations. We hope that researchers around the world can collaborate to improve it. But more importantly, we hope that it spurs innovation in many more areas."
After it won the television game show Jeopardy, IBM's Watson became one of the most well-known AIs in the world. IBM has continued to expand on Watson's capabilities and today it offers a host of services under the Watson brand name. IBM's focus is on providing cognitive computing (the company's preferred term for artificial intelligence) capabilities to developers so that they can incorporate them into their own applications. The Built with Watson section of its website features lots of examples of apps created using the technology, and it's quite possible that you have used more than one app that is powered by Watson technology. And of course, if you are a developer, it's also quite possible that you have used the Watson APIs yourself.
Purchased by Google in 2014 and now part of the Alphabet parent company , DeepMind is a London-based organization focused on artificial intelligence research. It made headlines last March, when its AlphaGo program became the first AI to beat the world's best players of the board game Go. If you play Go online, you may have unwittingly matched up against the AI. Recently, news broke that AlphaGo had been secretly participating in online tournaments under the pseudonym "Master" or "Magister." It won 50 matches in a row, sparking rampant speculation on Reddit about the mysterious player's identity. The DeepMind CEO later confessed that the company had been doing some "unofficial testing" of a new prototype AI.
You almost certainly have not used Viv today, but you might use it later this year. Viv is an artificial intelligence startup that Samsung acquired in late 2016. The Korean firm later announced that it would be including a new AI digital assistant in the Galaxy X8 smartphone, which is due for launch this year. That new digital assistant will be based on Viv technology. Samsung hasn't released a lot of details about the assistant, but has said that third-party developers will be able to attach and upload services that will work with the AI.
Einstein is Salesforce.com's AI, which is incorporated into its cloud-based enterprise software. So if your company uses Salesforce's Sales cloud, Service Cloud, Marketing Cloud or App Cloud, you may very well have interacted with this AI today. The company touts Einstein as "AI for everyone," and its focus is on making predictive analytics, machine learning and other AI services easy for businesspeople without a lot of technical expertise to use. It also offers Einstein capabilities to developers who are building IoT apps or enterprise apps that use Salesforce data. For those interested in learning more about artificial intelligence in general, its website has a wealth of easy-to-understand information.
Chatbots are becoming the go-to method for companies to provide basic customer service functionality, like answering simple questions. In many cases, the end users don't even realize they are talking to a machine rather than a human being.
Pandorabots claims to be the "world's leading chatbot platform" and provides the underlying technology for more than 285,000 chatbots. It offers a free integrated development environment for building chatbots, as well as paid services designed to make the process of building and deploying a bot faster and easier. It also integrates with third-party services like Twitter, Twilio and Slack.
19. Wolfram Alpha
If you have a question about almost anything, type it into the Wolfram Alpha website, and you're very likely to get a correct answer in response. This project's ambitious goal is "to make all systematic knowledge immediately computable and accessible to everyone." It already includes more than 10 trillion pieces of primary source data that is constantly being updated, and it has algorithms related to more than a thousand domains, as well as being able to perform more than fifty thousand kinds of equations and algorithms. Many other services, such as Microsoft Bing Search, DuckDuckGo search, Siri, Samsung S Voice and others, use Wolfram Alpha's capabilities to respond to some natural language inquiries from users.
Today artificial intelligence isn't just powering websites — in some cases, it's also designing those websites. Last year, the Wix website creation service launched a new capability called Wix ADI (short for "artificial design intelligence") that uses artificial intelligence to design custom websites. These aren't just template-based sites; the company claims that no two sites created with the service are the same. There's a demo video on the website that shows how it works. The service is currently available, so it's possible that you've already visited a website created by an AI without even realizing it.