A group of scientists from Tel Aviv University and Stanford University that is cooperating in the Digital Living 2030 program is trying to prepare for all of the scenarios in the coming years that will challenge the way we live. Prof. Irad Ben-Gal, who initiated the program and heads it together with Prof. Nick Bambos, says that while we have become accustomed to having a digital representation of our physical presence, which enables us to do things that are impossible in the physical world (such as sharing information simultaneously with thousands of people), people still think of themselves as physical beings. He believes that this is about to change.
"What will increasingly happen is that we will become dual beings fully living simultaneously in two worlds - the physical and the digital," says Ben-Gal. "Your digital being will be completely independent in making decisions. It is obvious that the digital being will make some decisions in the commercial world completely independently. Your digital being or agent can receive a budget and buy things for you or for itself. For example, it can order food for you when your refrigerator is empty."
"It can do other things. In the modern workplace, it seems completely natural to all of us for a company to put all of its knowledge into software that acts independently, but this principle can reach the level of an individual person. For example, a doctor can have a being that learns from the decisions of other doctors, and in time, will be able to give feedback, train the physical being, and examine other variables in order to make a decision. Later, while the physical doctor is sleeping, the digital doctor can go on working in his or her place. If the digital doctor gets enough information and creates value, it will be able to make decisions alone."
You can't control everything
When people assign more and more elements of their daily lives to smart digital beings that rely on huge databases, it is certainly possible that a person could be sent for an operation, chemotherapy, or a series of radiation treatments without ever seeing a physical human being. If this sound absurd, keep in mind that only last month, it was reported that a terminal patient from California discovered that he would die within a few days from a doctor whose face appeared on a video chat screen attached to a medical robot that entered his room.
While everyone agrees that creative and proper use of technology improves people's lives in a thousand and one ways, there is also general agreement that the accelerated technological development in recent decades poses new challenges to human beings. For this reason, in order to create on the one hand systems for optimizing the world's food supply or wiser electricity consumption, and on the other hand prevent cases in which the fate of human beings is decided by software or robots unemotionally notifying people that their end is near, it is necessary to understanding the sort of future towards which we are advancing. Research of the kind that Ben-Gal and his colleagues in the program hope to lead can spot the areas in which special problems will arise, and the researchers can then think how to respond to and prepare for them.
Digital Living 2030 connects scholars and students from Tel Aviv University and Stanford University for the development of infrastructure, procedures, methods, and algorithms carried out by hardware and software elements in order to support this new world.
Ben-Gal realizes that he is working in a world bustling with private entrepreneurship that does not stop to consider the deeper significance of accelerated technological development. Nevertheless, he regards higher education institutions as a player striving to channel things in a positive direction. "There is a realization that there are processes that cannot be managed from above, because they come independently from below. Still, there are ways, through regulation and a system of incentives, to control them," he says, referring to the effectiveness of higher education and the state in navigating the world of entrepreneurship. "You cannot really control startups, but you can guide them, for example, if you release information or data for their use in certain beneficial directions."
"Globes": Give an example.
Ben-Gal: "In transportation, for example, you provide access to all of the information pertaining to traffic lights, to geographic information systems, to traffic in real time, to the Internet of Things around the road, to municipality control rooms, and of course provide incentives for every municipality to take part in this. This will create a kind of infrastructure or matrix in which an independent startup can come with a specific solution, for example an algorithm for smart control of a traffic light, and then release its product that itself enters that infrastructure. Then another startup will take this product and improve it with the next stage of the system and so on."
Wisdom of the crowd as a hazard
Ben-Gal, 53, is a professor in the Tel Aviv University Faculty of Engineering and heads the AI and Machine Learning Laboratory there. He was a visiting professor at Stanford two years ago, and the similarity between the two universities - Stanford University's proximity to Silicon Valley and Tel Aviv University's proximity to the buoyant startup scene in Tel Aviv and Herzliya - caused him to initiate the program in partnership with Prof. Nick Bambos of Stanford. Prof. Michal Tzur, Prof. Oded Maimon, Prof. Joachim Meyer, Dr. Dan Yamin, Dr. Hila Chaluz Ben-Gal, Prof. Evgeni Khmelnitsky, and Dr. Erez Shmueli are also partners in the program. The program, which will continue for five years, received $5 million in support from the Koret Foundation, which supports joint research by scholars from Tel Aviv University and Stanford University.
The research itself is based on mathematical and engineering methodologies, such as operations research, artificial intelligence (AI), data science, as well as human factors and ergonomics. It deals with environmental, social, and personal developments that will affect the quality of our lives. Among other things, the research involves smart city apps, personalized and digital medical apps, decision-making systems based on AI agents, the future work environment, innovative supply chains, systems for preventing digital discrimination, and physical-digital systems of logic.
Adoption of new technologies has always had an enormous physical and mental effect on human development. Paleontologists believe that the use of fire shortened the time used to chew our food because of the ability to cook, resulting in a major increase in the size of our brains. The ability to produce basic stone tools enabled prehistoric mankind to overcome the fact that evolution had atrophied the sharp teeth and nails that other animals have. The invention of boats, wheels, cars, airplanes, and spaceships enabled people to travel ever-greater distances, thereby mixing gene pools that were previously isolated from each other.
The introduction of the printing press, followed later by radio, telephones, television, and the Internet, have given people the ability to communicate with each other in growing numbers in shorter timespans than ever before. The results of this communications revolution vary along a broad spectrum between the ability to stay in touch with relatives and do business in other places in the world and the risks created by the dissemination of falsehoods and incitement designed to influence millions of people. When all of this information is processed using AI, it is already clear that reliance on wisdom of the crowd, the assumption that the average of the answers to a question will converge on the correct result, does not necessarily serve our interests.
"We all realize that if you go to an expert, say a doctor, and show him or her a medical file, more weight should be given to a doctor with expertise in your specific problem," Ben-Gal explains. "One of the studies we conducted shows that if you have a group of specialists, each of whom is a specialist in one field, meaning that most of them are not specialists in your specific problem, working according to the wisdom of the crowd principle will consistently give you wrong answers. It is therefore frequently necessary to detect patterns in answers from the experts without knowing the right answer in advance. It will be difficult to find the experts in your particular problem in order to assign them more weight.
"The idea in this research is to find out to what extent experts are consistent in their joint decision-making when you present them with specific cases, for example medical cases. If we are both specialists in blood cancer, meaning that in all of the cases that show us a clear finding of blood cancer, we will think alike, while the answers by others who are less expert in this disease will be spread randomly among the other possible diseases. This means that the index of the proximity between us can be turned into the score of our expertise for the various cases, and will later be given more weight in the final opinion for the purpose of making a diagnosis."
Fairness of algorithms
Another sphere addressed in the program is the ethical consequences of a future in which smart systems guide our lives. "For example, imagine that you want to develop infrastructure in a country in the most efficient way using AI. You obtain a lot of information from Internet of Things networks that know where there are potholes in the road, where there is the possibility of accidents, and where traffic is dense. You can program an optimal algorithm to tell you the next interchange or road section that should be built, but this engineering optimization usually has no index of fairness or compassion - for example, prioritizing infrastructure fairly between neighborhoods."
The moral element
"Right. Therefore, in developing future AI algorithms, we will also have to incorporate these factors into the algorithm, and in addition to classical mathematical and engineering tools of operations research, also develop measures of fairness. You have to make a decision that is implanted in the algorithm - that if the most efficient solution fails to meet some kind of basic threshold of fairness to the population or person, it will not be used. This decision is critical, because once the decision passes to AI systems working ad hoc and very quickly, like a traffic light system, say, we have to take these aspects into account in advance."
Is there room for chance to play a role?
"It's possible. For example, in the way autonomous cars drive, the question arises whether it is preferable to run over a child or an old person, to collide with another car, or to drive off a cliff and kill the people in the vehicle itself. One of the ways to solve such a problem may be to leave elements of uncertainty in the algorithm. We accept uncertainty in people, and we have no ethical problem with a driver who collided with a tree and killed a passenger next to him, because that is what he decided at that moment. On the other hand, we have a problem if someone programs the algorithm in advance to respond in a specific way to a specific situation, such as the decision with what or whom to collide. The solution in a case of this kind may be to let the algorithm select the solution randomly. It is also possible that religious people might find divine providence in an AI algorithm."
Problems of a different sort arise from the possibility of allowing a digital being to study our patterns of behavior and analyze our views on a range of issues. Ostensibly, such a situation will enable people to make far more decisions on a broad range of subjects, and to create systems that will bring about democratization on a completely different level than we are familiar with. If you like, this is a far more advanced version of the existing practice of authorizing the politicians we elect to the Knesset to act in our names for a term in office. On the other hand, these systems will also constitute a target for other algorithms seeking to trick them.
In addition, the deep penetration of technology in our lives exacts a heavy personal price from us: addiction, attention and concentration deficiency, and damage to eyes and neck. "In this project, we are trying to address not only the prodigious opportunities that such a future offers organizations and society, but also the personal aspect - how much better people's lives will be in 2030, 2040, and after. We already know now that in many cases, this progress has had a negative impact on the quality of our lives. There is an inexhaustible load of information we are unable to bear, because we have limitations that digital systems do not have. As a result, we cannot maintain personal ties with many people simultaneously. The digital world therefore has an effect on the quality of connections between people. This new story that the digital world is opening up to us is forcing us to develop new systems, rules, and insights to cope with the new situation," Ben-Gal says.
The machines do not threaten me
Ben-Gal believes that we will soon see major changes in spheres such as autonomous transportation, digital and personalized medicine, smart cities, robots and AI-intensive industry, virtual environments, and applications that will affect the personal lives of all of us. On the personal level, he says, "We will see more complete integration between our digital world and our physical world. People will live simultaneously in the two worlds, with their digital being represented by digital agents performing many tasks for them: learning, making decisions in cooperation with other agents, conducting social interactions, etc."
At the same time, he believes that the basic needs of people as portrayed in Abraham Maslow's hierarchy of needs theory, which assumes that people have universal needs and will try to supply immediate needs before other needs, will remain similar. In such a world, physiological and digital needs will be at the bottom of the pyramid, the need for physical and digital security will be on the second level, the third level will be the need to belong, both physically and digitally, and above that will be the need for social appreciation in both physical and digital circles. The fifth level will be the need for self-fulfillment on through a physical and digital being.
Asked whether he believes that the future will be better than the reality we know now, Ben-Gal says that answer is subjective, and changes "from person to person, between different sections of society, between one region and another, and between one country and another."
For whom will it be better?
"Broad sections of society will enjoy a better life: personalized services, such as autonomous transportation based on need, and personalized health, a longer and healthier life, more enlightened use of resources, more leisure time, more efficient handling of the information burden, greater welfare as a result of streamlining of service and production processes, and a new and interesting range of professions. On the other hand, the digital revolution means new risks, such as risks to privacy and cybersecurity, the chance of more complicated and sophisticated crimes, social isolation of those who are not a party to progress, and side effects, such as widening gaps between wealthy groups and unsuccessful or less prosperous groups.
"First and foremost, there is a risk of widening economic and social gaps between people: between experts and non-experts in new areas, between rich and poor, between developed countries and developing countries, between technologically advanced sectors and those that are not technologically advanced. I think that this can reach very dangerous levels. As progress enables you to better take advantage of inefficiency, the risk of gaps grows, and if you are not capable of neutralizing this gap, you expose yourself to big risks. On the other hand, I am an optimist. As time passes, people are living longer, taking better care of their health, and can do a better job of dealing with inefficiency."
Asked to imagine what our lives will look like 110 years from now, Ben-Gal says, "Without taking the risk of justifying the saying that prophecy has been granted to fools, I will say that I am confident about one thing - the interaction between the digital world and the physical world will be complete; they will be inseparable. A person will be not merely a physical being with representation in digital worlds (such as our representation today on social networks); he or she will be a completely dual being. The digital being will have awareness, make independent decisions, connect with other beings professionally and socially, and encompass ranges of feelings and awareness that do not exist today.
"With the help of the digital being, each of us will be able to transcend the limitations that physics places on our physical beings. We will be able to be in several places at the same time, study without limitation, have several experiences simultaneously, take advantage of a free hour to take a virtual trip in virgin forests or the depths of the sea, connect with thousands of other beings professionally and personally, and participate in thousands of decisions that will greatly change the political situation by weakening the power of politicians and the role of the state as a dominant entity."
What about the popular apocalyptic scenario in which all of these smart beings develop awareness and rise up against us?
"That does not worry me. Machines lack many things they would need in order to threaten us in the way we are familiar with from science fiction. What does worry me is what will happen on a social level. It is clear to me that the young generation's experience of life will be very, very different. Human interactions will be completely different. When I look at more amorphous things, such as self-realization, satisfaction, belonging, and joie de vivre, I ask myself how these will appear in the new world"
Published by Globes, Israel business news - en.globes.co.il - on May 7, 2019
© Copyright of Globes Publisher Itonut (1983) Ltd. 2019