BetweenBrains

Taking back our AI Future

Omar Hatamleh, George Tilesch

• Publisher: GTPublishDrive
• Year: 2020
• eBook ISBN: 9781734931808
• Topic: Social Sciences
• Subtopic: Popular Culture
• Index: Social Sciences

AI is a present reality: we live at the threshold of an AI-dominated era. AI is more than a Technology wave. In the 21st century and beyond it is the very core source of power that fuels politics, business, and society: our minds, our work, and our homes. AI also rewrites the rules and forces us to lose no time in rethinking fundamental questions of our humanity. Our human brains will soon be required to adapt at hyper speed to a new paradigm of omnipresent machine intelligence.

We all have to deal with both its opportunities and threats in a conscious manner. Presently AI obscurity, hypotheticals, hype, and hysteria are aggravating problems of increasingly polarized and disconnected societies. However, whose image AI development will take and how the AI Age will be shaped is still in the hands of informed and clear-sighted citizens and leaders. The authors undertook the task of making sense of AI, especially its impact on the present and the near future for responsible readers worldwide.

CHAPTER 1
–
ON ROBONOMICS AND THE FUTURE OF HUMANITY

As children, we are all asked, “What do you want to be when you grow up?” We think of different things at different times, depending on what excites us, what is fashionable, what might impress other people. Perhaps you went on to do exactly what you hoped you would – but did it happen in exactly the way you expected? Imagining the future is not easy, and in some ways in this book we are asking what humanity wants to be when it grows up. We see some crucial choices coming, and some decisions that have already been made on our behalf. We want individuals, families, companies, societies, and states to take an active role in what comes next. Why? Because the new wave of change enabled by Artificial Intelligence (AI) will have a profound impact on society, our jobs, our health, and our politics.

What are the challenges and opportunities AI brings? As we will see, some are being addressed at different levels – by companies wanting to make money, by researchers looking for better solutions for everyday problems, by countries and organizations hoping to maintain society’s values in a parallel digital world. But there is more than just AI to take into account. For example, until now, no significant discussions have taken place regarding how population increases, and exponential technological advancement taken together have created a perfect storm for a shift and disruption to our future society and economic models. The narrative in the following chapters will touch on just a few of the seismic effects that AI could bring today and in the coming decades.

Do you remember your day-to-day life when you first started school or started working? How many of the tools you use every day now existed then? Now think what a child born in 2020 will see in their lifetime. It’s hard to imagine what the world will look like in 20 or 50 years, given the unprecedented acceleration in technological development and adoption, especially in the fields of data science and computational power. In stark contrast, if we turn the clock back 2,000 years, a prediction of what the world would look like in 50, 100, or even 1,000 years from that time would have been a relatively simple task.

Take for example, preeminent Greek philosopher and scientist Aristotle, who lived 2,400 years ago. His philosophy exerted a unique influence on almost every form of knowledge and his introduction of the method of scientific thinking is still relevant today. He had an especially brilliant mind but was working in a time when change was slower. If we consider the state of technological maturity of those times, Aristotle may have been able to imagine – and even adapt to – the way of living in the world 1,300 years after his time, when another great polymath, Ibn Sina (Avicenna), lived. Ibn Sina was regarded as one of the most influential physicians, thinkers, and astronomers of his time, and considered by many to be the father of modern medicine. We think Ibn Sina would have been able to envision and adapt to life in the 15th century, had he suddenly found himself a few hundred years into the future.

Perhaps the visionary who was most ahead of his time in predicting and designing early technological concepts several centuries ahead of his era was Leonardo da Vinci. The Italian polymath was born in 1452 and has long been regarded as the model Renaissance man and a prolific inventor. Leonardo was revered for his technological ingenuity. He conceptualized a bewildering number of inventions including technologies like flying machines, and armored fighting vehicles. Da Vinci may have also designed what may have been the very first humanoid robot by mobilizing a mechanized armored knight using gears and wheels connected to a pulley that enabled the basic “robot” to perform limited movements. However, due to the lack of scientific technology and advanced materials, his inventions were not feasible during that era, and his technological concepts remained unmaterialized. We might not be as creative as these individuals, but we are going to need to start thinking as they did to adapt to the pace of change that is coming.

We have advanced so much as a modern civilization that it is difficult to make sense of the fact that, just 100 years ago, Western society used horses as the primary means of transportation. Labor-saving machines brought radical change to factories and homes in the 19th and early 20th century, and in recent decades computers have been having a similar influence.

Unprecedented from the previous 2,000 years of human technological evolution, the present times are characterized by rapid, non-linear improvements in enabling technologies such as AI and Machine Learning (ML). ML is a branch of artificial intelligence that is based on the premise that the system can learn from data.

Computing machines date back to the abacus and took a leap forward with Charles Babbage’s Difference Engine and Analytical Engine. But it is really only in the last 90 years that modern computers were born and grew up with astonishing speed. Computer chips are increasing in power while the cost is diminishing. This is attributed to the fact that the number of transistors on a computer chip has been doubling roughly every 18 months. AI enthusiasts like Ray Kurzweil believe that the cognitive ability of a computer will soon be equivalent to that of a human. This will enable a computer to achieve complex tasks, such as performing medical diagnoses, creating music, or arguing with an opponent in a political debate. We need to prepare now for the moment when computers may become our equals, and maybe even outstrip our achievements.

What are the social and economic implications of such a paradigm shift in technological capabilities? In a few decades from now, if robots take on the form of humans, it may be hard to distinguish a humanoid robot from a human. Now imagine the possibilities and the societal implications of having advanced humanoid robots. What if they become as technically and cognitively capable as humans (or even superior to them) in several economically and socially relevant tasks and contexts? These technologies and emerging capabilities will significantly impact our society and the job markets of the future.

The establishment of shared theoretical frameworks, combined with the availability of data and processing power, has yielded remarkable successes in various component tasks such as speech recognition, image classification, autonomous vehicles, machine translation, and question-answering systems.

“As capabilities in these areas and others cross the threshold from laboratory research to economically valuable technologies, a virtuous cycle takes hold whereby even small improvements in performance are worth large sums of money, prompting greater investments in research. There is now a broad consensus that AI research is progressing steadily, and that its impact on society is likely to increase […]. Because of the great potential of AI, it is important to research how to reap its benefits while avoiding potential pitfalls.”

From an Open Letter, Research Priorities for Robust and Beneficial Artificial Intelligence, January 20151

There are discordant views about AI among scientists, techno-futurists and economists that are polarizing those in the general public. Some say that various forms of advanced robots and AI will create new and diverse job opportunities, while others argue that they will have devastating effects on the job markets of the future. Similar thoughts were circulated before the Industrial Revolution, when the majority believed that automation would devastate the job market; this turned out not to be the case.

The difference now is that we are not only competing for manual labor, but also for highly skilled jobs requiring intellectual skills and capability. Another important point to consider is that, for the first time in human history, value creation and human labor are separated.

Making Sense of AI

How will individuals and society as a whole evolve, and possibly (be) transform(ed) by these emerging and potentially pervasive technological capabilities? Should we expect co-evolution and cooperation between humans and AI-empowered humanoids or instead competition and potential conflicts?

Some of the answers to these profound, multi-faceted, and daunting questions will likely depend on if, when, and mostly how AI will reach levels of complexity, autonomous reasoning, and behavior(s) to de facto result in full AI capabilities of self-determination.

At this infant stage of AI evolution, humans are still playing a central and paradigmatic role in how they can conceive, design and shape the way of thinking of AI, from processing data to making sense of complex information and possibly acquiring real knowledge. In the future, autonomous robots driven by AI may even have a collective intellect and learning capabilities as they may be connected to each other. Whatever one robot learns, so do all others. That will substantially increase intellect and accelerate the rate of cognition in a way that is unparalleled to anything we have seen to date. What role will humans have in a not-too-distant future compared with intelligent, capable, and autonomous systems?

This is something of an existential challenge to humans and has sparked many concerns, hopes, and hypes surrounding AI. It’s crucial for the general public to clearly understand the distinctions between different levels of sophistication of AI to neither underestimate nor overestimate its capabilities. This will enable individuals and society to better assess the potential threats and benefits brought forward by AI in the future.

AI originally referred to the attempt to emulate a human-like form of intelligence, in all or at least most of its nuances. The term “AI” is often used in reference to powerful tools for data classification. These tools are impressive, but for now, they are on a totally different spectrum than human cognition. One of the fundamental challenges in emulating human intelligence is related to the fact that despite our extraordinary advances we do not fully understand how the human brain works. Consequently, we are not aware of any software, complex set of algorithms or AI platform that is even remotely as intelligent as the most cognitively challenged human being. Strictly speaking, no one has built (yet) anything close to real human-like AI capabilities on Earth.

At the current stage, AI is categorized as Artificial Narrow Intelligence (ANI or Narrow AI), which is limited in scope and restricted to only one functional area. ANI has the ability to outperform humans at some narrowly defined job. Two more stages – as yet theoretical – have been defined. The second stage of AI is Artificial General Intelligence (AGI). AGI also known as human-level AI can understand and reason in its environment as humans do and covers several fields that are on a par with adults. The third stage is Artificial Super Intelligence (ASI) in which AI surpasses human intelligence across practically all fields. Later on, we will see how industry, policy and society may be affected by each of these stages, and consider what actions might be sensible, or indeed valuable, to take now.

The quest for mimicry of intelligence has inspired the imagination of the general public for several decades. AI has been discussed for decades so why has it suddenly become one of the pervasive technologies that everyone is talking about? That is attributed to three main pillars.

First Pillar – the evolution of the fundamentals of neural networks (NN) and their associated complex algorithms. Early models for NN were single-layered and very basic, unable to perform complex cognitive tasks. Recent advancements in the field have resulted in more complex multi-layered NN that are significantly superior to the early models. These advanced algorithms are better suited to transcribing speech, and identifying objects including handwritten digits. Also, these advances in NN made it easier for the algorithms to recognize objects by looking at them from different angles.

Second Pillar – the video gaming industry had a significant impact by developing superior hardware to perform faster calculations by running calculations serially, thus allowing for more computations to occur simultaneously. It is astounding that in the 1960s when the integrated circuit (IC) was invented, an average IC contained two transistors. Today’s most sophisticated ICs comprise over 20 billion transistors.

Conventional Central Processing Units (CPUs) in PCs such as those manufactured by companies like Intel have been designed to run calculations serially. The rise of dual and quad-core CPUs has enabled the capability of the CPUs to perform more computations to occur simultaneously. However, current Graphics Processing Units (GPU) can contain several thousand cores in comparison, allowing them to process substantial amounts of data at the same time.

The CEO of NVIDIA predicts that GPUs will even get a thousand times faster by 2025 compared with the current ones. This will result in substantial benefits spanning across industries like healthcare, energy, cybersecurity, automotive, and financial services. The computational capabilities of parallel processors have advanced NN and machine learning and have substantially increased the processing speed for current existing AI algorithms, opening up various possibilities to expand their use across multiple platforms and technologies.

Third Pillar – availability of Big Data. AI requires data to build its capability, in particular, for machine learning. For example, a machine learning image recognition system, evaluates thousands of images of an object to understand and recognize later on. Therefore, Big Data derived from reliable and bias-free datasets can deliver the information required to train the algorithms to perform a superior function. AI algorithms continue to absorb new information and...

Table of contents

• COVER
• COPYRIGHT
• DEDICATION
• THE AUTHORS
• TABLE OF CONTENTS
• PROLOGUE TO THE FIRST EDITION
• CHAPTER 1 - ON ROBONOMICS AND THE FUTURE OF HUMANITY
• CHAPTER 2 - APPLIED AI: INDUSTRIES REWIRED
• CHAPTER 3 - AI - IN WHOSE IMAGE?
• CHAPTER 4 - THE GLOBAL RACE FOR AI DOMINANCE
• CHAPTER 5 - THE PATH OF THORNS: AI ETHICAL CHALLENGES
• CHAPTER 6 - AI & DEMOCRACY: IN MACHINES WE TRUST?
• CHAPTER 7 - FUTURE CASE STUDIES
• CHAPTER 8 - MANIFESTO - Taking back the Future: A new framework for ethical AI Stewardship
• EPILOGUE
• FURTHER READING
• BETWEENBRAINS - TAKING BACK OUR FUTURE IN THE AI AGE
• NOTES
• BACKCOVER