We talk about artificial intelligence (AI), robots, and machine learning as if they’re coming soon, or are just some tech pipe dream.
They’re not. They’re here today.
In fact, a special report from Bank of America, Merrill Lynch predicts the global market for AI and robots will be just under $153 billion by 2020, and some industries will experience up to a 30% productivity increase through the use of those technologies alone.
That’s not a century from now; it’s not even a decade. It’s just three short years away. That can either terrify you if you’ve seen too many sci-fi films, or excite you if you consider the upside and benefits it could yield.
The reality probably lies somewhere in the middle. Positives and negatives, good and bad. There will be disruption – there will be jobs and perhaps even whole industries that see massive displacement from robots and other “intelligent” machines.
And that says nothing of the inherent risk associated with creating something capable of logical thinking without emotion. The robots may not rise up and exterminate humanity any time soon, but the development of true AI is closer than you think. We already have it to varying degrees. And while computer scientists haven’t yet created a truly sentient artificial being, their work in the field is already having tremendous impacts on several industries.
Artificial Intelligence vs. Machine Learning
Before we take a look at some of the ways it’s changing the world around us, let’s make clear the difference between two key components.
Artificial intelligence is defined as a computer program capable of performing tasks that usually require human intelligence, such as speech recognition, translation from one language to another, or decision making. It involves taking steps or changing course in order to achieve some kind of goal.
Machine learning is a type of AI where computer systems can actually learn, improve, and “evolve” when exposed to new and additional data. They don’t need to be programmed in the traditional sense. They discern new information using existing knowledge, make connections, combine ideas, and following a train of thought just as humans do.
“You essentially have software writing software.” – Jen-Hsun Huang, CEO of Nvidia
Machine learning falls under the broad umbrella of AI. Similar, but different; both are transforming the landscape in amazing ways.
Hotly contested and debated by those in and outside of the industry, the idea of robots replacing teachers seems, too many, ludicrous. After all, teachers explain more than just maths and science. They’re responsible for modeling socialization skills, proper behavior, self-esteem, confidence, self-regulation, and a host of other life skills not covered in a textbook.
Can a robot do all of that? Doubtful, but it remains to be seen.
What we do know, though, is that technology has already moved into the classroom in a big way, with interactive smartboards, laptop computers, wifi, tablets, voting devices, multimedia presentations, and so on.
Some tech professionals believe that AI could further improve the student experience by providing differentiation on a scale not possible with a human teacher. Each student would receive an individual learning plan – designed by a program using machine learning – that addresses their unique learning style and needs.
AI and machine learning would allow for faster identification of learning disabilities. Student work could be collected, compiled, analyzed, and compared against known difficulties. Diagnosing a learning obstacle can currently take years, and is often missed completely.
Some teachers have turned to AI to generate tests and assignments. A program can sift through thousands of sources and pick the key questions and vocabulary that best demonstrate comprehension of any given subject.
Intelligent tutor systems (ITS) like AutoTutor can actually evaluate student steps and problem solving to identify misconceptions and gaps in students’ knowledge. They can guide, give hints, and provide feedback on a student’s performance.
Online learning has been broadened by machine learning systems that not only allow teachers to reach students far away – possibly beyond the reach of a traditional classroom – but also work to highlight their weaknesses and areas that need attention. A teacher can efficiently “teach” a larger student body without sacrificing quality or attention.
Outside of the traditional classroom, language programs like Carnegie Speech and Duolingo harness the power of automatic speech recognition (ASR) and natural language processing (NLP) to listen for speaking errors and correct them with users. With them, students can master even a native accent without ever having a native instructor.
Anyone can learn a language anywhere, no tutor or language partner required. Other AI programs exist for a wide variety of skills and topics. The line between formal, classroom-based education and self-study is blurred like never before.
AI could benefit education in countless ways, but we’re more reluctant to allow “robots” to instruct our kids than we are to let them drive our cars.
And that’s too bad. Learning analytics could evaluate and improve curricula and develop a better student experience. We could use adaptive learning systems like DreamBox and recommendation engines to meet the individual needs of all students.
Denying that we live in a tech and digital world does a tremendous disservice to students. It’s time to fully embrace the potential – not as a replacement for teachers, but as a 21st century complement to them.
You may be a bit hesitant to trust your health to a computer system, but chances are good that you already have (at least, in partnership with your doctor).
Diagnosis, prescription, treatment, and even medical procedures themselves are getting the machine learning enhancement seen almost everywhere else.
Computers can quickly compare symptoms and genetic details against millions of possible diagnoses, determining the probability of each and the most likely cause in a fraction of the time it takes a human doctor.
New biometric sensors and massive databases allow machines to process, analyze, and compare data that could lead to faster diagnoses for many diseases and conditions.
Faster diagnosis means a faster recovery – or even the difference between life and death.
This is not Googling your symptoms. It’s not self-diagnosing. Robo-doctors are the machine equivalent of a trained professional, but with a high-tech twist.
Beyond that, robot surgeons can perform operations with an unwavering “hand,” near-perfect precision, and virtually zero possibility of error. That means smaller surgery scars, faster healing, less time in hospitals, fewer (if any) mistakes, and ultimately, less expense. That’s a major win for patients.
Intuitive Surgical Inc. has a robotic laparoscopic surgery system called da Vinci that performed nearly 600,000 procedures in 2014 alone. It’s not a replacement for a surgeon (yet), but rather a robotic extension of his or her skills.
Of course, robots lack the human touch – the bedside manner – of flesh-and-blood doctors, and that may be a massive strike against them.
“It’s lethal to think that you can separate the psychological care from the physical care.” – Richard Lilford, University of Warwick’s Chairman of Public Health
IDC predicts that up to 30% of care providers will be using cognitive analytics to assist with patient data by 2018. That’s almost a third of healthcare workers turning to AI for help and suggestions within the next 12 months.
These systems could also potentially identify individuals at greater risk of developing diseases and illnesses based on genetic background, lifestyle, and other big data available to the “doctor-bots.”
As the saying goes, an ounce of prevention is worth a pound of cure.
Lumiata, for example, is bringing AI to predictive medicine. Its motto? “Health can be unpredictable. We’re using AI to change that.”
Lumiata’s program simultaneously looks at patient records, family history, ongoing studies, medical journals and articles, and other unstructured data sources to make educated predictions about patient health in the short and long-term.
It combines data and medical science, which basically means it learns in much the same way a doctor does.
And it doesn’t stop after treatment. Some practitioners are using AI bots like NextIT as a virtual doctor-on-call with patients after they’ve gone home. It can answer complex questions, send reminders about medications and recommendations, and even ask about symptoms and side effects. It then relays everything back to the doctor.
Diagnostics and imaging. Patient monitoring. Drug discovery. Treatment. Follow up. Industry advancements. New techniques. There are no fewer than 106 startups working with AI in the healthcare field today, covering everything across the full spectrum.
Healthcare is getting a 21st-century upgrade. Robots may never replace doctors entirely, but they’re already lending a steel-plated hand.
Ready to take a seat in a self-driving car?
Waymo (formerly the Google self-driving car project) promises to eliminate tired, distracted, and drunk driving, making the roads safer for everyone. It’s already accumulated the equivalent of 300 years of driving experience on city streets since 2009. A complex system of sensors and software keeps the car informed of everything happening around it.
Uber has developed a fleet of self-driving Volvos (the two companies are spending $300 million to further develop them) operating in Pittsburgh and Arizona. They still require human operators in the car to take over operation if necessary, and the program got some bad publicity when the cars were caught driving through red lights in California (the company lost its permit to use them in the state shortly thereafter), but Uber is pushing forward with the initiative.
Tesla cars come equipped with its new Autopilot feature, a system of eight cameras and 12 sensors that provide a 360-degree view around the vehicle to a range of 250 meters. In addition to “seeing” the car in front of you, Autopilot can detect environmental conditions like rain, fog, and dust, and react accordingly.
Besides the usual functions – maintaining speed, staying in the lane, slowing down as necessary – Autopilot can actually change lanes by itself, transition from one highway to another, exit the freeway, self-park, and even be summoned to and from your garage!
Gartner predicts upwards of 250 million smart cars on the roads by 2020, connected to high-tech networks, each other, and the environment around them.
They’ll offer advanced driver assistance systems (ADAS) including driver condition analysis, camera vision systems, engine control, detection units, and more.
If you remove the human, the possibility of human error disappears. Will it make the roads safer? That’s the working theory.
In the transport industry, Daimler Trucks recently unveiled its Freightliner Inspection Truck, an 18-wheel and fully autonomous transport truck licensed to drive on American roadways. It’s not driverless in the strictest sense of the word, but it is capable of taking over under certain conditions. Once engaged, it can maintain the speed limit, keep a safe distance from other vehicles and objects, remain in its lane, and slow down or stop when necessary.
Will it mark the end of overworked and exhausted long-haul truckers?
And what about the aviation industry? Is there a place for AI and machine learning on your flight from New York to London? Absolutely. In fact, it’s already there.
The autopilot system is primarily responsible for flying the plane between takeoff and landing (although most systems are able to land by themselves if required). That’s not to suggest the human pilot is unnecessary; far from it. He or she monitors the system and takes over as necessary.
But their biggest role is probably providing a sense of comfort for the passengers themselves. A driverless car is one thing, but how many would be willing to fly 13 hours without a pilot?
We may not be far from that happening. Many – though not all – airline crashes are attributed to human error, and short of rare malfunctions or catastrophic breakdowns, computers just don’t make mistakes. It could be argued that a pilotless airplane is the safest airplane.
We already have driverless trains, subways, and mass transit systems. We’re inching closer to a time when a driver of any sort will be nothing but a memory.
4. Financial Services
AI and machine learning have taken hold in the financial services arena in a big way. Just under a third of respondents in a recent survey confirmed using the technology for voice recognition and response, recommendation engines, predictive analytics, and more.
Many banks are using complex algorithms to assess loan risk, and approve or deny based on their conclusion alone. The traditional loan officer is no longer needed, other than to pass along the decision to the client.
Robo-advisors are set to disrupt the industry in a way never before seen, and there are already plenty of them plying their trade online. Some estimates have them controlling about $2 trillion in assets by 2020. They could be managing 17% of total assets within the next five years, according to others.
So what’s the appeal? Well, these robo-advisors can quickly sort through thousands of investment opportunities and stocks to match them with an individual’s profile, including risk tolerance, long-term goals, and so forth.
Machine learning systems evaluate your online presence for clues about your values, beliefs, and personal ethics to further align your portfolio with everything about you. The system can make real-time portfolio adjustments based on market fluctuations and changes in personal circumstances.
And let’s not forget about the powerful prediction and probability algorithms that are used to make decisions about this stock over that stock, or the best time to buy or sell. AI and machine learning software can process thousands – if not millions – of data points in the blink of an eye. You’ll never miss a once-in-a-lifetime opportunity again.
It doesn’t stop there. Intelligent virtual assistants and service reps can help customers with a wide variety of questions and problems. The national bank of Sweden recently deployed Nina, an online assistant that achieved 30,000 conversations per month, and first-contact resolution of 78%.
Want to ask your bank account a question? You can with Clinc’s Finie app. It provides intelligent financial advice and answers for your day-to-day living.
You can ask it how much you spent on dining out this month, and whether it was more or less than last month. You can check whether you can afford to take a vacation this winter, or how much you have left for groceries in your weekly budget, or almost anything else related to your account, your expenses, and your budget.
How about automatic fraud detection? Machine learning platforms like Feedzai monitor millions of data points and transactions for unusual behavior and spending patterns. When it notices something out of the ordinary, it raises a red flag and is quickly pulled up for review or blocked entirely.
Risk assessment. Systems acting as guarantors for loans. Online trading platforms. Self-help. The list goes on. The potential for AI to take over – for better or worse – is huge.
5. Business and Marketing
General business is not immune to the AI invasion.
An MIT survey of 168 large companies found that 76% are using machine learning technologies to assist their sales growth strategies.
MarketMuse is banking on AI taking over your content marketing strategy, too. Its product uses AI and machine learning to determine the best topics to write about, and how to cover them completely. It then analyzes your existing content inventory to create a comprehensive content plan designed specifically for you and your website. It helps you find gaps, makes suggestions, and keeps you timely and relevant.
Other use cases include recommendation engines like those seen on Amazon, the New York Times, and Netflix. These learning algorithms examine your past activity or purchases, and compare them against millions of others to (fairly) accurately predict what you might like to buy or read or watch next.
Google’s own RankBrain uses AI and learning to better process search queries and return more relevant results. It gets smarter with each inquiry.
A service like SailThru apply machine learning to your email marketing. It analyzes customer behavior to identify the perfect time to connect and engage, delivering double-digit increases in revenue for its clients.
Intelligent chatbots to communicate and assist customers 24/7. Incredible data-backed insights and customer profiles. Increased productivity. Higher profits. When it comes to AI and ML in business, the sky’s the limit.
Want to try it for yourself? Import.io enables users to convert the mass of data on websites into structured, machine readable data with no coding required. Request a free trial here.
AI and machine learning are popping up everywhere and for everything.
Heck, even the new iPhone 7 includes machine learning in its top-notch camera, which will take better pictures over time. A swarm AI platform known as UNU correctly picked the top four finishers at the Kentucky Derby. The NFL uses it to analyze the movement of its players to reorganize play style.
“New developments in machine intelligence will make us far, far smarter as a result, for everyone on the planet.” – Eric Schmidt, Executive Chairman at Google.
The solar industry. The entertainment industry. Your Roomba vacuum cleaner; those things are seriously high-tech, using an AI mix of navigation, self-charging, problem solving, and even modeling systems that allow it to map your house and develop more efficient methods and routes for cleaning it.
So what industries are being changed by machine learning and artificial intelligence? A better question is, which ones aren’t? (It’s a much shorter list.)
Sing up for Import.io and start getting web data to feed your machine learning and artificial intelligence projects.