The Rise of Artificial Intelligence: AI and Robotics
Section 1: Introduction and History
Beware the Fourth Industrial Revolution! The Robots are coming! The Robots are coming! Do we need a modern-day Paul Revere to call the country to arms? Maybe not just yet. . .
The rise of Artificial Intelligence (AI) and Robotics from 1970 to today has been persistent, amazing, and both a benefit and challenge to mankind. Although we did not achieve Marvin Minsky’s 1970 prediction, that by the end of the decade we would have a computer with the general intelligence of a human (Wadhwa 1), there have been fantastic and unanticipated strides. As with any technological change, there were supporters, early adopters, and starry- eyed evangelists, as well as, detractors, non-believers, and doomsayers. But the progress of AI and Robotics continued forward throughout those years at a steady pace.
1970’s Initial Steps and Mixed Expectations
This decade began the era of real steps in using AI and Robotics to assist mankind and solve greater problems than just industrial robots on the manufacturing assembly line. At Waseda University in Japan, the first anthropomorphic robot, the WALBOT-1, a biped, sighted, conversational robot was built (Press, 7). At Carnegie Mellon University, the XCON eXpert CONfiguer is created which used rule-based logic to automatically customize the configuration of DEC VAX computers. The Stanford Cart is created, an autonomous vehicle that successfully and independently crosses a room full of chairs in five hours (Press, 7). All these initial steps advance the reality that AI and Robotics can move, see, communicate, and create. But at the same time, James Lighthill of the British Science Research Council states that “in the field of AI no discoveries made so far produced the major impact as promised’ (Smith, 18).
1980’s A Change in Direction
In the early 80’s the Japanese Ministry of International Trade and Industry budgeted $850 million for the Fifth Generation Computer Project. These funds were targeted to develop computers to communicate, translate language, interpret pictures, and apply human-like reasoning (Press 8). Mercedes-Benz built the first driverless car which traveled at 55 mph on closed streets. Researchers from the University of California and Carnegie Mellon publish a new learning procedure based on back-propagation for computer networks of ‘neuron like’ configuration (Rumelhart, 4). At IBM’s Watson Research Center, a statistical approach to language translation proposes a change from rule-based Expert System intelligence to machine learning based on statistical analysis (Press 9).
1990’s Data Analysis Maturity and Neural Network Puberty
As the timeline progressed, greater effort was placed on developing neural networks that extended beyond brute-force data analysis and proposed to build intelligent systems that learn from interaction with the world around them. A.L.I.C.E. was created which leveraged natural language data collection to expand the chatbot’s abilities (Press 10). Furby, the first pet robot was created. And IBM’s Deep Blue, used custom Very Large-Scale Integration chips and parallel processing to defeat the reigning chess champion (Wadhwa 1).
2000’s Machine Learning and Applications
As the industry and science continued, applications developed which used increased computing power, neural processing, and other advances in robotics to create more advance anthropomorphic designs: MIT develops Kismet, a machine that can mimic and recognize emotions, and ASIMO, the Honda robot, can walk as fast as a human (Press 10). This along with advances in multi-layer neural networks and graphics processors allow machines to develop strong top-down connections to take steps to achieve deep learning using data representations instead of task specific rules.
2010’s Neural Networks Mature
Today, we now have very large neural network machines that approach and, in some cases, exceed human thinking capability. A neural network designed at the German Traffic Sign Recognition completion achieved a 99.46% accuracy versus 99.22% for humans (Press 11). IBM’s Watson defeated two human champions at the game show Jeopardy!, and Google’s DeepMind AlphaGo defeated Lee Sedol, the professional Go player of 9 dan rank (Press 12). These achievements in game playing and visual recognition may seem trivial, but they foretell the impact AI and Robotics evolution.
Section 2: Recent Evolution
In recent years, the rapid evolution of AI and Robotics has been driven by machine learning more than by brute force analytics and calculations. New neural network techniques patterned after human learning allows machines to process information in layers and form connections between the layers. As these layers and connections increase or mature, the machine becomes more capable to do human things: differentially respond to new inputs, act based on environmental cues, perform highly complex physical maneuvers and recognize image, voice, and text.
This deep learning enables robots to perform repetitive, hazardous, or speed intensive work, but also identify and act on different courses of action in real-time. AI and Robotics, once the domain of factories and universities, has now begun to spread out into our home, work, health, and social life. The technology is becoming productized. When products get created and people begin to appreciate the benefits and pay for their acquisition, the power of capitalism comes into play. The economic engine will fund the proliferation of AI applications and put an economical robot at the hands of the many, not just corporations or the very wealthy (Baker-Louie 3).
These AI and robotic applications are no longer the domain of computer science. Profit motivated companies are now able to leverage the technology to create products in many business and services. One example in the use of AI is Google’s DeepMind. This is a Google-owned company whose business model is to combine advanced techniques from machine learning and systems neuroscience to structure intelligence and implement it into machines and understand the human brain. An example of an application of this model is the collaboration of DeepMind and the Moorsfield Eye Hospital (Cag 1). This partnership is focused on the diagnosis and treatment of macular degeneration. An age-related vision disorder that relies on eye scans which take significant time for humans to analyze and diagnose. DeepMind is helping to automatically diagnose these scans and begin treatment programs immediately. Saving days and weeks of untreated vision degeneration.
Another example of the business application of AI is for companies that have accumulated large data sources of customer purchases, complaints, complements, socioeconomic status, buying patterns, and scores of other attributes. AI driven data analysis allows companies to consume this data and determine patterns that would take countless hours of human brain power to deduce. A partnership using IBM’s Watson platform and Fluid, a digital retail company, is targeted to create products and services for client companies to leverage and identify highly-relevant product recommendations to their customers (Cag 3). AI creates the ‘white matter’ which harnesses the ‘gray matter’ of accumulated data using machine learning to think like the ultimate Marketing Manager.
Finally, another significant example of the continued evolution of AI and Robotics, is the impact on customer service and the replacement of the person using automated switchboards to directly communicate with clients. AI programs not only respond to customer questions, but also differently react to human responses to learn and interact, effectively replacing the human agent. IPsoft’s Amelia, is an example of such a ‘software robot’ in use in the British public sector in Enfield Council, London (Cag 4). Enfield is one of London’s largest boroughs and is growing by thousands of constituents each year. Amelia receives 100,000 visits to its website and takes 55,000 telephone calls each month. Allowing consistency and accuracy in customer service to meet demands on the government.
Section 3: Impacts on the Human Experience
These advances create significant improvements to healthcare, the workplace, entertainment, our quality of life, and society in general. But this reliance on AI and Robotics also has effects on employment opportunities, reduces career advancement in corporations, creates new and different types of mental and social stress, and causes health issues unanticipated even a few years ago.
Intelligent robots reduce human jobs and wages as fewer humans and manufacturing facilities are required to produce products and keep up with consumer demands. Entry level positions now require more training and education as the price of employment. There is increasingly less opportunity to start at the bottom and work your way up the corporate ladder and acquire knowledge by initially working at less complicated tasks and growing your skills year over year. Greater screen time for humans has led to diminished health and more obesity. Sitting in front of a screen all day replaces smoking as the biggest health threat. Sitting is becoming the new cancer.
In July 2016 in Dallas, police used a robot for the first time to apply lethal force to stop an armed suspect after five police officers were murdered (Cag 5). During both Middle East wars in recent years, US Military forces used AI to determine targets to attack based on input from flying drones and employed robots to clear minefields and place bombs in humanly inaccessible locations. This use of technology has protected humans, however the main objective of AI means that a machine determines the best action and takes it. This may ultimately make it impossible to predict the behavior of AI and robotic driven weapon systems.
However, we must also appreciate the many positive aspects of this technology. AI enhances human efficiency and performance. Humans are freed from the tedious, mundane tasks we can use our robot proxies to perform. This creates freedom for humans to focus on innovation, imagination, less physical and more intellectually rewarding occupations. New technology can mean more jobs, not less jobs. Just as the steam engine, the automotive, and the computer did not merely eliminate workers, but created new industries and improved our way of living. New AI and robotic based industries will demand knowledge workers to design, build, employ, and manage this technology across medicine, business, manufacturing, communications, and all the diverse areas of human existence.
Robotic support for minimally invasive surgery, diagnosis of diseases and disorders, and replacement of human limbs and organs are perhaps some of the best example of the benefits of the human-machine interface. Robotic protheses, wheelchairs, and other devices help persons with disabilities in all areas of life. AI supported chemical and genetic research means the world’s best doctors, laboratories, and hospitals can produce medicines and cures at a rate never known. Those cures can be replicated across the world multiplying the positive impact to Earth’s human experience.
These benefits elevate our lives, protect our loved ones, provide us with joy, entertainment, and fascination. When asked if there is any moment in history that someone would want to live, the impact of AI and robotics on the human experience could force you to choose today rather than any other period.
Section 4: Acceptance and Transformation
The challenge for humankind today is determine how to take advantage of the positive effects of AI and Robotics while not experiencing the negative effects on our health, family life, and professional careers. As human ‘masters’ we must control the rights and responsibilities that govern technology. Artificial Intelligence and Robotics are here to stay. Just like airplanes, televisions, and cell phones. Personal assistants will continue to keep getting smarter and our reliance on them will keep getting stronger. Embedded AI technology continues to make existing technology faster, more capable, easier to use, and more reliable. As AI technology matures and becomes ever more integrated and connected with other technology, we will live in a Matrix-like world of machines, automation, software intelligence, and androids.
Embracing the technology is the path forward. However, we cannot absolve humans of responsibility for machine actions. Just as we do not absolve the driver of a car from the results of an automobile accident. We must respect Steve Hawking’s warning. “The development of full artificial intelligence could spell the end of the human race” (Cag 6), but not be intimidated by it.
Section 5: Conclusion
What do we do? First, we need to ensure there is regulatory control at both the national and international levels to ensure we protect our human rights and liberties. Unregulated technology historically has led to diminished respect for the human condition, pursuit of wealth at all costs, and human pain and misery. We need to look no further than the history of the textile mills in England, the coal mines and railroads in the United States, the slave ships leaving Africa providing labor for cotton gins, and more examples too numerous to list. Regulation with real oversight and control is a must.
Next, we need to fully fund and support training of everyone to live in the world of the future. Technology advancement does not mean lack of a need for human talents. James Bessen documents how the manufacturing labor force grew from less than 12% in 1820 to 26% by 1920 despite advances in automation (Willige 2). In our era, computer-based jobs can replace other occupations. Alternative Currency Bankers, Global System Architects, Avatar Designers, 3D Printing Engineers can replace the call center and factory workers of today. It will require education, internships, and mentorships, but these are the jobs of the future and the future is what we create.
Finally, humans will need to have a new mindset. One based on leveraging the truly human qualities of exploration, curiosity, innovation, dreams, desires, hopes, human values, and free will. We are on an infinite journey before machines can ever acquire those skills. If they ever will. Those skills were a gift to humans from a deity a long, long time ago. A gift only that deity can bestow.