In ‘The Terminator’ series of action films starring Arnold Schwarzenegger, a cybernetic organism (cyborg) is programmed from the future to go back in time and kill the mother of the scientist who leads the fight against Skynet, an artificial intelligence system that will cause a nuclear holocaust.
Terrifying and at times comical (“I’ll be back”, “Make my day”) The Terminator cyborg was among the first presentations of artificial intelligence (AI) to a global audience. (“C3PO” and “R2D2” of Star Wars fame and “Data” from Star Trek also qualify) The robotic assassin also depicted an extreme scenario of what could happen when “machine learning”, a vital element of AI, runs amok – ie. when machine intelligence surpasses human intelligence to the point when the machines take over and try to eliminate the humans who created them.
While numerous facets of AI have been developed over the past couple of decades, all with positive outcomes, the fear of AI being programmed to do something devastating to the human race, of computers “going rogue”, continues to persist.
On the other hand, AI holds tremendous potential for benefiting humanity in ways we are only just starting to recognize. This article gives an overview of artificial intelligence including some of its most interesting manifestations.
What is AI?
The first step is defining what we mean by artificial intelligence.
One definition of AI is “the simulation of human intelligence processes by machines, especially computers.” Such processes include learning by acquiring information, understanding the rules around using that information, employing reasoning to reach conclusions, and self-correcting.
Another definition understands artificial intelligence as “a machine’s ability to understand, think and act on a problem in the same way a human would in the same situation. But there is no limit placed on the form the AI will take. AI can manifest itself invisibly. Far away from the human eye, working on advanced problems, it can take the form of a self-driving car, parts of a factory, or even in the future some sort of advanced robotics. What is important to understand is that there is no limit on the forms AI can or will take, now or in the future, to interact with us.”
Optimistically, AI is seen as a new industrial revolution, with the unprecedented ability to unlock the constraints of the human mind. Perceived as one of the most interesting and powerful innovations that will shape the next 100 years, AI’s potential for radical industrial and social transformation is on par with the invention of the steam engine, electricity, and the personal computer.
A computer is thought to be AI-capable when, presented with a task, it can find a solution without human intervention. It learns as it goes. Two landmark examples are the 2016 victory of a computer in the complex Chinese board game “Go” versus the game’s world champion (there are said to be as many moves in Go as atoms in the universe); and when a computer, playing hundreds of games of Atari’s “Breakout”, played the game better than humans, even figuring out a creative way of winning, by digging a tunnel that helped it efficiently knock out rows of blocks.
12x growth potential
Before getting into its applications, a brief market analysis of AI reveals some startling numbers from an investment perspective. The global market for artificial intelligence was valued at US$20.67 billion in 2018, with North America accounting for nearly half, $9.72 billion. A recent report from Fortune Insights has the North American market growing at a blistering 33.1% CAGR to 2026, with global AI revenues soaring 12-fold, to around $118 billion. In junior resource investing, that’s what we call a 12-bagger.
Growth areas are expected to be in natural language processing (programming computers to analyze human language), robotic process automation, and machine learning. Major drivers also include the industrial Internet of Things (IoT), higher usage of big data and robotics in manufacturing, significantly more venture capital investment, and cross-industry partnerships.
Another report, published last month, says the growth of AI in manufacturing will be influenced by the largest players in the market including NVIDIA, Oracle, Microsoft, Intel and IBM. The Research and Markets report states that North America will dominate the AI manufacturing market until 2024, but that Asia Pacific will be the fastest growing region, due to increasing investments in AI technologies.
US vs China
Central to this dynamic is China’s push to lead the world in artificial intelligence. In 2017 Beijing stated its intention to dominate the sector by 2030. China has rolled out facial recognition technology on a wide scale, prompting many observers to claim the Chinese Politburo is using AI to extend the long arm of the state through widespread surveillance.
“On AI, China is implementing the technology very fast in facial recognition, speed recognition, self-driving vehicles, smart cities and medical diagnoses,” Rebecca Fannin, author of ‘Tech Titans of China,’ told CNBC.
The country is already the leader in financial technology, well ahead of North America. Chinese consumers pay for purchases on their phones – through apps like WeChat – or facial recognition software. Need a loan? An application made via smart phone can be processed within seconds, using software that evaluates hundreds of variables including the ability to pay based on body language and cell phone habits. “Cash is dead in China. So are credit cards,” says Fannin.
Many believe that over time, the ever-expanding artificial intelligence industry will be split between the US and China, with increasing animosity between the two superpowers as they jostle for global AI and tech supremacy.
Computer science professor Tom Mitchell, who founded the world’s first research center for artificial intelligence at Carnegie Mellon, believes that, while the US has more experience building tech companies, China has a clear advantage in using big data sets. He uses the the medical field as an example. The US has had an electronic medical records system for over 20 years but it has yet to employ machine-learning algorithms, due to privacy concerns and for-profit medical care. “In China, it’s a different situation. If the government decides that it’s going to have country-wide electronic medical records… then it’s going to happen,” he told BBC News.
Sticking with the medical theme, advancements in AI could be a powerful force for good, particularly with respect to early cancer detection.
The ability to identify breast cancer early is a critical factor in determining how a patient will respond to treatment. Researchers are beginning to use AI to diagnose cancerous lesions before it’s too late. At MIT’s Artificial Intelligence Lab in Boston, Professor Regina Barzilay is using deep learning to teach a computer how to understand language, and read text and data. Accessing thousands of images from the Breast Imaging Center at Massachusetts General Hospital (MGH), Prof. Barzilay and Dr. Connie Lehman, head of the Breast Imaging Center, amassed a large data set of all the patients who had had breast surgery, for a certain type of high-risk lesion. With machine learning, they could separate out the patients who needed surgery, from those who could avoid it.
The predictive outcome of artificial intelligence was astounding. The AI-enabled computer could read hundreds of thousands of images where the outcome is known, learn the properties of the cancer-spot images, such as how the pixels are distributed, and then determine, what are the patterns that correlate highly with future occurrences of breast cancer.
“So instead of using human capacity to recognize patterns, which is inherently limited by our cognitive capacity, and how much we can see and remember, we’re providing the machine with a lot of data and making it learn this prediction,” says Prof. Barzilay, in a recent Frontline documentary titled ‘In the Age of AI’.
“It’s going to change the face of breast cancer,” adds Dr. Lehman. “We are using this technology not only to be better at assessing the breast density, but to get more to the point of what we’re trying to predict. Does this woman have a cancer now, and will she develop a cancer in five years? That’s where deep learning and artificial intelligence can really help us in our patients.”
Digital Science reported in 2018, a study showing a computer could detect melanoma nearly 10% more accurately than dermatologists. Shown over 100,000 images, the deep-learning computer diagnosed skin cancer with 95% accuracy, compared to 86% accuracy from 58 dermatologists shown the same images.
An even higher level of accuracy was found in a 2015 study that used Artificial Immune System technology to detect mesothelioma, a cancer affecting the thin layer of tissue covering the lungs, frequently caused by exposure to asbestos fibers.
The New York Genome Center relies on an artificial intelligence system by IBM’s Watson supercomputer, for screening its patients for glioblastoma, a form of brain cancer.
Finally, Chinese start-up Infervision is using image recognition technology and deep learning to diagnose signs of lung cancer with X-rays.
Robots that care
Artificial intelligence contained within the sterile, inhuman shell of a robot is a concept that has been with us since the 1950s. For decades robots have been thought of as rather dumb, simple machines that talk in monotone. They’ve even found their way into our language. “Robotic” refers to someone who performs a task mechanically, or is devoid of personality.
Factory workers fear that robots will replace them, and to some extent, they have. US manufacturing has become heavily automated. It’s very unlikely that we will ever see a return to the middle class being paid high wages to do repetitive tasks. Widespread factory automation is here to stay.
Lately however, research is developing to turn the job-killing robot into an emotionally-intelligent machine that can actually deliver care to those who need it most. With advances in medicine, people are living longer. As Baby Boomers get into their golden years and Generation Xers begin to retire, the need for at-home and long-term care is likely to outstrip the number of health care workers available.
A mechanical doctor or assistant would come in useful in cases where constant monitoring and follow-up is needed, such as recovering from a heart attack or stroke.
A mobile robot called ENRICHME is designed to help older people with a range of tasks that become challenging later in life.
Nova reports, The robot was tested in retirement homes in three European countries to see if it could help combat cognitive decline and improve quality of life. Early results show that the users accepted the robot, that it helped them be both more cognitively and physically active, as well as solving some difficulties they meet in everyday life, like finding missing items.
Mabu, a smiling yellow robot holding a touch screen, reminds residents at care homes to take their medications. Aifloo, a Swedish company that develops e-health solutions based on AI, is working on a system for tracking nursing home patients.
AI can also be put to use detecting strokes and heart attacks. An organization called Cambio Health Care developed a product that alerts a physician when a patient is at risk of having a stroke. Coala Life’s heart monitor quickly analyzes heart sounds and rhythms to detect atrial fibrillation and other arrhythmias and heart murmurs.
Robots are also slowly being incorporated into classrooms.
An expressive classroom robot called Milo, developed by robotics company RoboKind, says Milo “never gets tired, never gets frustrated and is always consistent”. The robotic ability to repeat something again and again in the same tone without tiring is particularly suited to helping children with ASD [Autism Spectrum Disorder]learn, states Nova, adding that Milo is also showing benefits for kids with Down Syndrome, ADHD and other social or learning problems.
Driverless vehicles don’t make mistakes. Or do they?
Automated vehicles are often pointed to as the poster child for artificial intelligence, with great strides having been made in recent years to dispatch driver-less vehicles to easily-driven routes.
In 2016 Uber made history for making the first automated truck delivery, 50,000 cans of beer from a brewery in Fort Collins to Colorado Springs, 120 miles away.
Wired reminds us that, in December 2018, Waymo, the product of Google’s self-driving car project, officially started its commercial self-driving car service in the suburbs of Phoenix. Its monopoly is being challenged by smaller start-ups like May Mobility and Drive.ai. Ride-hailing companies Lyft and Uber are on board with autonomous driving, as are tech giants like Apple, IBM and Intel.
While many believe the shift from human- to robot-piloted vehicles is inevitable, with the substantial benefits of adding $7 trillion to the global economy and saving thousands of lives, others are skeptical. The collapse of unionized trucking in the US has whittled a truck driver’s income from what used to be a decent wage, to barely scraping by. Autonomous vehicles are likely to be the final nail in the coffin of the once-might Teamsters. Other victims are likely to include big chunks of the auto industry, gas stations, drive-thrus and taxi drivers.
As for being the safer alternative, in March 2018 the first death by an autonomous vehicle was recorded in Tempe, Arizona. A 49-year-old woman was struck at 39 mph by a fully-automated, self-driving Uber car, as she was walking her bicycle across the road in the dark. In its investigation, the Transportation Safety Board revealed that the car couldn’t recognize her as a pedestrian because she wasn’t crossing at a cross-walk.
Autonomous vehicles apparently aren’t programmed to identify jaywalkers. In my opinion, there’s still a long ways to go before driverless vehicles become a common sight on North American roads and highways.
Making life better
More likely to inhabit our personal spaces now, and in the near future, are very task-specific AI inventions that, while not necessary for daily living, make life just a little bit better. Some of these have been with us for a while.
Robotic vacuums such as iRobot’s popular Roomba clean your house, remember the layout to clean more efficiently, dump their own dirt in a receptacle, and even find their way back to the recharging station. Siri and Alexa, the ubiquitous voices installed on your smart phone, may know your comings and goings better than your spouse. The revolutionary voice-recognition programs help us to quickly scan the web, shop and schedule appointments. They also help power and heat smart homes, and assist those with mobility issues.
Cogito, Boxever and John Paul are examples of AI programs that improve both customer service and a client’s shopping experience. While not generally recognized as AI, major shopping portals like Amazon, Netflix and Pandora are also employing machine learning to identify a user’s tastes and suggest products, movies or songs, based on the user’s search history.
AI for evil
Of course, artificial intelligence is not all about positive outcomes. There are some decidedly negative consequences of allowing machines to do the work human beings used to do. Two of the best examples are AI’s ability to conduct Orwellian-like video surveillance, and AI-powered drones that can kill or maim with clinical precision.
It’s estimated there are about 500,000 CCTV cameras mounted in London, and 4 to 6 million throughout the United Kingdom. Londoners would probably be less okay with CCTV if they knew the power of today’s surveillance technology.
Before artificial intelligence, surveillance cameras had limited use. The computers at the other end being monitored by a bored security guard were only capable of storing the footage, usually for a limited time before getting deleted. As Vice reports,
Increasingly, none of that is true. Recent developments in video analytics—fueled by artificial intelligence techniques like machine learning—enable computers to watch and understand surveillance videos with human-like discernment. Identification technologies make it easier to automatically figure out who is in the videos. And finally, the cameras themselves have become cheaper, more ubiquitous, and much better; cameras mounted on drones can effectively watch an entire city. Computers can watch all the video without human issues like distraction, fatigue, training, or needing to be paid. The result is a level of surveillance that was impossible just a few years ago.
An ACLU report called “the Dawn of Robot Surveillance” says AI-aided video surveillance “won’t just record us, but will also make judgments about us based on their understanding of our actions, emotions, skin colour, clothing, voice, and more. These automated ‘video analytics’ technologies threaten to fundamentally change the nature of surveillance.”
Nowadays, the computers connected to the cameras can tell whether someone is walking in the right direction, they can count people and cars, identify people based on their clothing, analyze their body language, and even tell a customer’s sentiment, such as Amazon’s in-store cameras. The technology already exists for cameras to automatically identify someone, in real time. As Vice presciently notes, Facial recognition technology is improving all the time, made easier by the enormous stockpile of tagged photographs we give to Facebook and other social media sites, and the photos governments collect in the process of issuing ID cards and drivers licenses.
Think your online identity is private? Police already have the ability to track cell phones, and identify people from information broadcast by smart phones, laptops and Bluetooth-connected devices.
Western governments deem such power to be in the interests of national security and/ or crime prevention, but as we all know from the case of Edward Snowden, the former CIA employee/ subcontractor and whistle blower who copied and leaked highly classified information from the National Security Agency, it can also be used to spy on citizens.
Among the most disturbing uses of AI for social control and suppressing political activity is what is currently happening in China. Starting next month, telecom carriers must scan the faces of anyone applying for mobile or Internet service. Is there any doubt the Chinese government will make good use of this valuable information?
Quartz reports that on October 4, the Hong Kong government invoked emergency powers outlawing demonstrators from wearing face masks. Rioters who are identified face up to 10 years in prison, or worse in Mainland China.
There, facial recognition software and AI have combined to produce a “social credit score” for every citizen that helps authorities to reward or punish behaviour – even jaywalking. This program that in the West would be considered abhorrent, a gross invasion of privacy, is reportedly quite popular among Chinese citizens.
A 2019 report by the Carnegie Endowment for International Peace finds a growing number of countries are deploying AI surveillance tools to surveil citizens for “a range of policy objectives – some lawful, others that violate human rights, and many of which fall into a murky middle ground.” The report, via Al Jazeera, says at least 75 of 176 countries are actively using AI for surveillance purposes. China is the largest supplier of these technologies.
The use of drones, (or UAVs, short for unmanned aerial vehicles) is probably the most frightening manifestation of AI, for their ability to be used in first-strike attacks and futuristic machine-to-machine warfare. Drones under the control of a human operator can target and kill enemies from thousands of miles away.
In September Houthi-backed rebels in Yemen claimed responsibility for drone strikes on two oil facilities in Saudi Arabia including Aramco’s largest oil processing plant.
The Pentagon recently warned that China is exporting “next-generation drones” to countries in the Middle East – who under a weapons embargo are banned from purchasing advanced American UAVs. The graphic below shows China’s drone fleet has a ceiling of between 4 and 9 kilometers, and a payload from 80 to 1,200 kilograms. China’s killer drones are most frequently seen flying the skies above Saudi Arabia, Jordan, Nigeria, Yemen, Iraq, and the UAE, writes Zero Hedge.
Of course, drones are being flown in plenty of other scenarios, some extremely useful. For instance, a group of Virginia Tech engineers is hoping to employ drones in search and rescue missions: Using mathematical models based on historical data that reflect what lost people actually do combined with typical searcher behavior, the researchers hope this novel approach of balancing autonomy with human collaboration can make searches more effective, Eureka reports.
Unmanned vehicles are also being used to increasingly greater effect for scanning and mapping building sites; helping to solve traffic problems; to drop supplies to disaster victims; for providing farmers with intel as to when is the best time to plant their crops, and how much fertilizer to use; for inspecting infrastructure like pipelines and power lines; surveying damaged buildings for insurance purposes; and by film-makers or news media to capture aerial footage.
From this relatively brief analysis, it should be apparent that artificial intelligence, despite being a force for good and evil, opens up a genie’s bottle worth of investable ideas. The AI “revolution” has only just started and it is moving forward at a rapid pace. We are talking about a 33% compound annual growth rate – a 12-bagger, in resource investing terms.
In particular, I like how AI shows the potential of being bolted onto just about every industry – including 3D printing which to me is a mini-revolution in itself, transforming the manufacture of everything from rockets to human organs.