On IBM’s Watson homepage, their mission statement is “Observe, interpret, evaluate, decide.” Since 2014, Watson has been leading the charge of what software can do in bridging the gap between big data and human understanding. Not only does Watson gather information, it's currently touted as an assistant to doctors for making the right healthcare decisions for patients, as it deals with the all-too-human idea of “gray areas.” Such advances could greatly advance medical care, not to mention a host of other industries, as we know it, freeing up doctors’ manpower to pursue treatment rather than spend time trying to find the perfect decision.
IBM is so invested in Watson that not only did they name the technology after IBM’s founder’s son Thomas Watson Jr., they have heralded it as a symbol of the “third era of computing.” CEO Ginni Rometty declared: "In the future, every decision that mankind makes is going to be informed by a cognitive system like Watson."
Using big data analysis to make decisions is nothing new. For years now, social networking and e-commerce sites have received some media attention and ire in the public as the depth of their data mining and selling has come to light, and there have been debates over how much privacy is too little. How many times have you shopped on Amazon only to see an ad later on for the items you were browsing on a different site?
But what if instead of a data visualization tool or logic-based software pouring over data to help companies make a decision, software could, instead, make the decisions on its own? And not only could it make that decision, but act on it with an emotional, human touch?
“A large part of your brain is shackled by the boredom and drudgery of everyday existence,” Chetan Dube, CEO and founder of IPsoft, told Entrepreneur magazine this month in “Even Better than the Real Thing.” “You have to drive a car, vacuum the floor or take the garbage out. But imagine if technology could come along and take care of all these mundane chores for you, and allow you to indulge in the forms of creative expression that only the human brain can indulge in. What a wonderful world we would be able to create around us.”
That’s the world Dube wants artificial intelligence (AI) to make possible. Today, AI such as IBM’s Watson deal with processing mountains of information to draw conclusions so that the humans utilizing the technology have more time to act on those resolutions and make decisions. In short: The artificial intelligence of 2015 has its roots firmly planted in big data.
For IPsoft, those piles of data were employee manuals and guidelines as well as the skills employees picked up on the job everyday. The Manhattan-based company, founded in 1998, built “remote infrastructure management solutions designed to automate the range of IT environments and business processes.” Their IPcenter program acted as a center of “virtual engineers” communicating with each other to “brainstorm” solutions to problems, using a company’s infrastructure of information to do so. From this program, "Amelia" was born.
In the article, Jason Ankeny describes her (yes, her) as a “virtual agent avatar poised to redefine how enterprises operate by automating and enhancing a wide range of business processes.” Amelia is a bit of software that not only appears human, but acts it as well, thanks to her cognitive learning software. Amelia can respond to queries as well as analyze intent and emotion to resolve issues quickly and effectively, all with a human touch thanks to her blue-eyed-blonde interface.
Dube describes her development process as that of simulating and emulating how thoughts are formed rather than attempting to “reverse-engineer” the brain. The (patent-pending) software in Amelia’s brain is able to create conceptual maps as she goes through each interaction -- a cumulative hub of every employee’s intellect. Amelia, who actually happens to resemble Dube's wife, relies on “baseline” information to reply to emails and calls, and when she doesn’t know an answer, or how to handle a situation, she’ll hand the issue off to a human representative -- and then learn based on that interaction for the next time such an issue is presented.
Amelia’s intelligence algorithms grasp the “deeper implications of what she reads, applying logic and making connections between concepts.” She can absorb the information in employee manuals and guidelines that would take mere humans months to master in a matter of seconds. She is able to comprehend and respond to natural language and complex issues.
And, she’s good at her job. In beta trials with one of their enterprise partners, Dube says that by the end of her first month on the job she was handling 42 percent of requests (a 320 percent increase from the 10 percent in her first week) and by the end of the second month she was taking on 61 percent of all incoming requests.
Artificial intelligence rooted in evolution
While AI behaving like humans sounds like a sci-fi apocalypse, the technology behind Amelia’s ability to analyze and respond to human emotion isn’t as frightening as it would appear. Cogito Corporation’s voice analytics, for example, is simply science.
"Our technology is based on the concept of 'honest signals' from evolutionary biology, or the idea that important and relevant communication of a person's internal state are activated by the primitive brain and conveyed through behavior and actions largely subconsciously, thus largely uncontrollable and 'honest,’” Ali Azarbayejani, Cogito’s CTO, said. What this means it that it’s not what you say, but how you say it, that gives the greatest context for communication.
“Imagine listening to a conversation between two people speaking a foreign language that you do not understand,” he said. “Despite not understanding the linguistic content, you nonetheless can usually tell a lot about the conversation. Are they angry? Are they in agreement or disagreement? Is one participant dominant or are they sharing the conversation?”
The development behind such software is a relatively straightforward concept, as well. Cogito uses audio signal processing, which then categorizes “vocal acoustic properties, prosody, and conversational interaction patterns.” Their program then maps the interaction based on metrics. They call their signals “Big Signals,” as in “big data signals.”
“Traditional data analysis assumes that all data is in some structured form, such as a database,” Azarbayejani said. “The whole traditional framework obviously is insufficient when most of the information we care about is embedded in audio signals, stored as audio files. Metrics or 'features' on audio signals can be extracted and treated as traditional data, but the set of possible metrics is effectively infinite.”
To confront the vast amount of signals, Cogito has developed a methodology and modeling toolset that takes all of the signals into account to “build perceptually relevant, theory-driven, customer-validated algorithms and statistical models.” Implemented in Java, they’ve built their stack using NoSQL databases, MongoDB in particular, for their “flexible sharding and scalability” properties. The service then uses C++ to offer real-time signal computation.
Gartner research forecasts that by 2017, AI like Cogito’s voice analytics and Amelia will fuel a 60 percent reduction in the cost of IT solutions by “automating repetitive tasks currently tackled by humans” -- and even end outsourcing as we know it, which is great news for anyone who has called a customer service line in the last few years or so. Cognitive AI can also make your job easier by taking over mundane tasks, which, depending on your industry, can comprise a significant portion of your average workday -- or, just take your job completely.
IBM has a quote from Thomas Watson Jr. on Watson’s site, which states:
“Our machines should be nothing more than tools for extending the powers of the human beings that use them.”
Even so, Watson, who doesn't even appear human, has been met with reservations. Business Insider questioned if IBM was, perhaps, just trying to find a problem for its solution. Lauren Friedman argued that to “make that future a reality, IBM has a huge task in front of it...It must convince companies...that ‘cognitive computing’ — where computers are involved in decision-making, not just data-crunching — is something they actually need.”
And while AI like Amelia may free up your time, she may also replace basic human functions, says Alex Kozlov, the director of content at Alsbridge. “What’s happening isn’t that you eliminate a 50-person department,” he told Entrepreneur. “Instead, you implement a smart tool, and each of those fifty people had thirty percent more time on their hands.” As a result, management may need to restructure the jobs of the individuals left, cut out a portion of the workforce, or create new tasks. This could lead to job loss on a massive scale.
One thing civilization needs in abundance is jobs. Management consulting firm McKinsey and Company predicts that by 2025, the 250 million flesh and blood humans currently performing tasks, particularly in the IT and customer service industry, will be taken over by AI in the form of automation technology -- bad news for anyone out there living paycheck to paycheck, up to their eyeballs in student loans, or relying on the benefits of a full-time job. In addition to job loss, could AI really mean the end of humanity as we know it?
Is it really the end?
Experts have long warned about AI and its possible negative side effects; just take a look at all of the dystopian sci-fi horror out there. As Ankeny noted, last October, Tesla founder Elon Musk compared AI to “summoning the demon. You know all those stories where there's the guy with the pentagram and the holy water and he's like... yeah, he's sure he can control the demon, [but] it doesn't work out." Likewise, Stephen Hawking recently warned the BBC that the “development of artificial intelligence could spell the end of the human race.”
To some extent, this is true. Innovation tends to be spurred by challenges humanity faces, but some of the greatest leaps in technology have been made before the public knew it was needed. Apple has staked its entire business model on this idea, toting around their wares that we’re not quite sure we need, or really want, as International Business Times has pointed out about the Apple Watch. After all, did you really know that you needed the iPod and, later, the iPhone, before it first came out?
Luckily, most current applications of AI are just accessories to humans in the workforce, enhancing our understanding and helping us make decisions based on what we already know without having to drudge through mountains of information, good news for everyone out there who is sick of the status quo. Quocirca estimates that the average IT staffer devotes thirty percent of his or her time to basic, repetitive tasks in the workplace. Imagine what you could do with that extra thirty percent (or two hours and forty-two minutes in nine hour workday) of time.