Artificial Intelligence and the Future of Humans

The use of AI reduces individuals’ control over their lives

Autonomous systems can reduce or eliminate the need for human involvement in some tasks. Today's ever-advancing artificial narrow intelligence (ANI) tools – for instance, search engines and digital "agents" such as Siri, Alexa and Cortana – are not close to reaching the goal of human-like artificial general intelligence (AGI). They are, however, continually becoming more powerful thanks to developments in machine learning and natural language processing and advances in materials science, networking, energy-storage and hardware capabilities.

ANI is machine intelligence that equals or exceeds people's abilities or efficiency at a specific task. For years, code-based tools in robots and other systems have performed repetitive tasks like factory-floor assembly activities. Today, these tools are quickly evolving to master human traits such as reason, logic, learning, task-performance and creativity. Today's smart, networked, software-equipped devices, cars, digital assistants and platforms, such as Google search and Facebook social mapping, accomplish extremely complex tasks. The systems underpinning today's global financial markets, businesses, militaries, police forces, and medical, energy and industrial operations are all dependent upon networked AI of one type or another.

What is the future of humans in an age of accelerating technological change?

Many experts in this canvassing said that as AI advances human autonomy and agency are at risk. They note that decision-making on key aspects of life is ceded to code-driven tools. Individuals who function in this digital world sacrifice, to varying degrees, their independence, right to privacy and power over choice. Many of the experts who worry about this say humans accede to this in order to stay competitive, to participate socially and professionally in the world, to be entertained and to get things done. They say people hand over some control of their lives because of the perceived advantages they gain via digital tools – efficiency, convenience and superior pattern recognition, data storage, and search-and-find capabilities. Here is a selection of responses from these experts that touch on this:

An anonymous respondent summed up the concerns of many, writing, "The most-feared reversal in human fortune of the AI age is loss of agency. The trade-off for the near-instant, low-friction convenience of digital life is the loss of context about and control over its processes. People's blind dependence on digital tools is deepening as automated systems become more complex and ownership of those systems is by the elite".

Baratunde Thurston, futurist, former director of digital at The Onion and co-founder of comedy/technology start-up Cultivated Wit, said, "For the record, this is not the future I want, but it is what I expect given existing default settings in our economic and sociopolitical system preferences. … The problems to which we are applying machine learning and AI are generally not ones that will lead to a 'better' life for most people. That's why I say in 2030, most people won't be better due to AI. We won't be more autonomous; we will be more automated as we follow the metaphorical GPS line through daily interactions. We don't choose our breakfast or our morning workouts or our route to work. An algorithm will make these choices for us in a way that maximizes efficiency (narrowly defined) and probably also maximizes the profitability of the service provider. By 2030, we may cram more activities and interactions into our days, but I don't think that will make our lives 'better'. A better life, by my definition, is one in which we feel more valued and happy. Given that the biggest investments in AI are on behalf of marketing efforts designed to deplete our attention and bank balances, I can only imagine this leading to days that are more filled but lives that are less fulfilled. To create a different future, I believe we must unleash these technologies toward goals beyond profit maximization. Imagine a mapping app that plotted your work commute through the most beautiful route, not simply the fastest. Imagine a communications app that facilitated deeper connections with people you deemed most important. These technologies must be more people-centric. We need to ask that they ask us, 'What is important to you? How would you like to spend your time?' But that's not the system we're building. All those decisions have been hoarded by the unimaginative pursuit of profit".

Thad Hall, a researcher and coauthor of "Politics for a Connected American Public," added: "AI is likely to have benefits – from improving medical diagnoses to improving people's consumer experiences. However, there are four aspects of AI that are very problematic. 1) It is likely to result in more economic uncertainty and dislocation for people, including employment issues and more need to change jobs to stay relevant. 2) AI will continue to erode people's privacy as search becomes more thorough. China's monitoring of populations illustrates what this could look like in authoritarian and Western countries, with greater facial recognition used to identify people and affect their privacy. 3) AI will likely continue to have biases that are negative toward minority populations, including groups we have not considered. Given that algorithms often have identifiable biases (e.g., favoring people who are white or male), they likely also have biases that are less well-recognized, such as biases that are negative toward people with disabilities, older people or other groups. These biases may ripple through society in unknown ways. Some groups are more likely to be monitored effectively. 4) AI is creating a world where reality can be manipulated in ways we do not appreciate. Fake videos, audio and similar media are likely to explode and create a world where 'reality' is hard to discern. The relativistic political world will become more so, with people having evidence to support their own reality or multiple realities that mean no one knows what is the 'truth'".

Thomas Schneider, head of International Relations Service and vice-director at the Federal Office of Communications (OFCOM) in Switzerland, said, "AI will help mankind to be more efficient, live safer and healthier, and manage resources like energy, transport, etc., more efficiently. At the same time, there are a number of risks that AI may be used by those in power to manipulate, control and dominate others. (We have seen this with every new technology: It can and will be used for good and bad). Much will depend about how AI will be governed: If we have an inclusive and bottom-up governance system of well-informed citizens, then AI will be used for improving our quality of life. If only a few people decide about how AI is used and what for, many others will be dependent on the decisions of these few and risk being manipulated by them. The biggest danger in my view is that there will be a greater pressure on all members of our societies to live according to what 'the system' will tell us is 'best for us' to do and not to do, i.e., that we may lose the autonomy to decide ourselves how we want to live our lives, to choose diverse ways of doing things. With more and more 'recommendations,' 'rankings' and competition through social pressure and control, we may risk a loss of individual fundamental freedoms (including but not limited to the right to a private life) that we have fought for in the last decades and centuries".

Bart Knijnenburg, assistant professor of computer science who is active in the Human Factors Institute at Clemson University, said, "Whether AI will make our lives better depends on how it is implemented. Many current AI systems (including adaptive content-presentation systems and so-called recommender systems) try to avoid information and choice overload by replacing our decision-making processes with algorithmic predictions. True empowerment will come from these systems supporting rather than replacing our decision-making practices. This is the only way we can overcome choice/information overload and at the same time avoid so-called 'filter bubbles'. For example, Facebook's current post ranking systems will eventually turn us all into cat video watching zombies, because they follow our behavioral patterns, which may not be aligned with our preferences. The algorithms behind these tools need to support human agency, not replace it".

Peter Reiner, professor and co-founder of the National Core for Neuroethics at the University of British Columbia, commented, "I am confident that in 2030 both arms of this query will be true: AI-driven algorithms will substantially enhance our abilities as humans and human autonomy and agency will be diminished. Whether people will be better off than they are today is a separate question, and the answer depends to a substantial degree on how looming technological developments unfold. On the one hand, if corporate entities retain unbridled control over how AI-driven algorithms interact with humans, people will be less well off, as the loss of autonomy and agency will be largely to the benefit of the corporations. On the other hand, if 'we the people' demand that corporate entities deploy AI-algorithms in a manner that is sensitive to the issues of human autonomy and agency, then there is a real possibility for us to be better off – enhanced by the power of the AI-driven algorithm and yet not relegated to an impoverished seat at the decision-making table. One could even parse this further, anticipating that certain decisions can be comfortably left in the hands of the AI-driven algorithm, with other decisions either falling back on humans or arrived at through a combination of AI-driven algorithmic input and human decision making. If we approach these issues skillfully – and it will take quite a bit of collaborative work between ethicists and industry – we can have the best of both worlds. On the other hand, if we are lax in acting as watchdogs over industry we will be functionally rich and decisionally poor".

Paul Vixie, an Internet Hall of Fame member known for designing and implementing several Domain Name System protocol extensions and applications, wrote, "Understanding is a perfect proxy for control. As we make more of the world's economy non-understandable by the masses, we make it easier for powerful interests to practice control. Real autonomy or privacy or unpredictability will be seen as a threat and managed around".

João Pedro Taveira, embedded systems researcher and smart grids architect for INOV INESC Inovação in Portugal, wrote, "Basically, we will lose several degrees of freedom. Are we ready for that? When we wake up to what is happening it might be too late to do anything about it. Artificial intelligence is a subject that must be studied philosophically, in open-minded, abstract and hypothetical ways. Using this perspective, the issues to be solved by humans are (but not limited to) AI, feelings, values, motivation, free will, solidarity, love and hate. Yes, we will have serious problems. Dropping the 'artificial' off AI, look at the concept of intelligence. As a computer-science person, I know that so-called 'AI' studies how an agent (a software program) increases its knowledge base using rules that are defined using pattern-recognition mechanisms. No matter which mechanisms are used to generate this rule set, the result will be always behavioral profiling. Right now, everybody uses and agrees to use a wide set of appliances, services and products without a full understanding of the information that is being shared with enterprises, companies and other parties. There's a lack of needed regulation and audit mechanisms on who or what uses our information and how it is used and whether it is stored for future use. Governments and others will try to access this information using these tools by decree, arguing national security or administration efficiency improvements. Enterprises and companies might argue that these tools offer improvement of quality of service, but there's no guarantee about individuals' privacy, anonymity, individual security, intractability and so on".

Ramon Lopez de Mantaras, director of the Spanish National Research Council's Artificial Intelligence Research Institute, said, "I do not think it is a good idea to give high levels of autonomy to AI systems. They are, and will be, weak AI systems without commonsense knowledge. They will have more and more competence, yes, but this will be competence without comprehension. AI machines should remain at the level of tools or, at most, assistants, always keeping the human in the loop. We should all read or re-read the book 'Computer Power and Human Reason' by Joseph Weizenbaum before deciding whether or not to give lots of autonomy to stupid machines".

Oscar Gandy, emeritus professor of communication at the University of Pennsylvania, responded, "AI systems will make quite substantial and important contributions to the ability of health care providers to generate accurate diagnoses of maladies and threats to my well-being, now and in the future. I can imagine the development and deployment of systems in which my well-being is the primary basis of our relationship. I am less sure about how my access to and use of this resource may be constrained or distorted by the interests of the other actors (humans within profit/power-seeking orientations). I assume that they will be aided by their own AI systems informing them how to best present options to me. I am hopeful that we will have agents (whether private, social, governmental) whose interest and responsibility is in ensuring that my interests govern those relationships".

Robert Epstein, senior research psychologist at the American Institute for Behavioral Research and Technology and the founding director of the Loebner Prize, a competition in artificial intelligence, said, "By 2030, it is likely that AIs will have achieved a type of sentience, even if it is not human-like. They will also be able to exercise varying degrees of control over most human communications, financial transactions, transportation systems, power grids and weapon systems. As I noted in my 2008 book, 'Parsing the Turing Test,' they will reside in the 'InterNest' we have been building for them, and we will have no way of dislodging them. How they decide to deal with humanity – to help us, ignore us or destroy us – will be entirely up to them, and there is no way currently to predict which avenue they will choose. Because a few paranoid humans will almost certainly try to destroy the new sentient AIs, there is at least a reasonable possibility that that they will swat us like the flies we are – the possibility that Stephen Hawking, Elon Musk and others have warned about. There is no way, to my knowledge, of stopping this future from emerging. Driven by the convenience of connectivity, the greed that underlies business expansion and the pipedreams of muddle-headed people who confuse machine-like intelligence with biological intelligence, we will continue to build AIs we can barely understand and to expand the InterNest in which they will live – until the inevitable – whatever that proves to be – occurs".

An attorney specializing in policy issues for a global digital rights organization commented, "I'm not sure, even today, whether the tech advances of the last 12 years have been net positive over the global population. We've seen a widening gap between the very rich and everybody else. That is likely bad for democracy. AI seems likely to make the employment/training problem worse in the U.S., and AI may have similar effects in countries that currently provide cheap labor. On the political-governmental side, AI will exacerbate current surveillance and accountability problems. I figure that AI will improve and speed up all biometric pattern recognition as well as DNA analysis and natural language processing. And though we know that much of this is biased, we're not adequately counteracting the bias we know about. The companies who generate and disseminate AI technology have every incentive to continue. I'm not optimistic that collective action – at least in the U.S. system – will successfully counter those incentives".

Brian Behlendorf, executive director of the Hyperledger project at The Linux Foundation and expert in blockchain technology, wrote, "I am concerned that AI will not be a democratizing power, but will enhance further the power and wealth of those who already hold it. This is because more data means better AI, and data is expensive to acquire, especially personal data, the most valuable kind. This is in contrast to networking technologies, whose benefits were shared fairly widely as the prices for components came down equally fast for everyone. One other reason: AI apps will be harder to debug than ordinary apps, and we already see hard-to-debug applications leading to disenfranchisement and deterioration of living. So, I do not take as a given that AI will enrich 'most' people's lives over the next 12 years".

Eileen Donahoe, executive director of the Global Digital Policy Incubator at Stanford University, commented, "While I do believe human-machines collaboration will bring many benefits to society over time, I fear that we will not have made enough progress by 2030 to ensure that benefits will be spread evenly or to protect against downside risks, especially as they relate to bias, discrimination and loss of accountability by that time".

David Bray, executive director of People-Centered Internet, commented, "Hope: Human-machine/AI collaborations extend our abilities of humans while we (humans) intentionally strive to preserve values of respect, dignity and agency of choice for individuals. Machines bring together different groups of people and communities and help us work and live together by reflecting on our own biases and helping us come to understand the plurality of different perspectives of others. Big concern: Human-machine/AI collaborations turn out to not benefit everyone, only a few, and result in a form of 'indentured servitude' or 'neo-feudalism' that is not people-centered and not uplifting of people. Machines amplify existing confirmation biases and other human characteristics resulting in sensationalist, emotion-ridden news and other communications that gets page views and ad-clicks yet lack nuance of understanding, resulting in tribalism and a devolution of open societies and pluralities to the detriment of the global human condition".

Bernie Hogan, senior research fellow at Oxford Internet Institute, wrote, "The current political and economic climate suggests that existing technology, especially machine learning, will be used to create better decisions for those in power while creating an ever more tedious morass of bureaucracy for the rest. We see little example of successful bottom-up technology, open source technology and hacktivism relative to the encroaching surveillance state and attention economy".

Dan Buehrer, a retired professor of computer science formerly with the National Chung Cheng University in Taiwan, warned, "Statistics will be replaced by individualized models, thus allowing control of all individuals by totalitarian states and, eventually, by socially intelligent machines".

Nathalie Marechal, doctoral candidate at the University of Southern California's Annenberg School for Communication who researches the intersection of internet policy and human rights, said, "Absent rapid and decisive actions to rein in both government overreach and companies' amoral quest for profit, technological developments – including AI – will bring about the infrastructure for total social control, threatening democracy and the right to individual self-determination".

Katja Grace, contributor to the AI Impacts research project and a research associate with the Machine Intelligence Research Institute, said, "There is a substantial chance that AI will leave everyone worse off, perhaps radically so. The chance is less than 50 percent, but the downside risk is so large that there could be an expectation the world might be worse for AI".

David A. Banks, an associate research analyst with the Social Science Research Council, said, "AI will be very useful to a small professional class but will be used to monitor and control everyone else".

Luis German Rodriguez Leal, teacher and researcher at the Universidad Central de Venezuela and consultant on technology for development, said, "Humankind is not addressing properly the issue of educating people about possibilities and risks of human-machine/AI collaboration. One can observe today the growing problems of ill-intentioned manipulation of information and technological resources. There are already plenty of examples about how decision-making is biased using big data, machine learning, privacy violations and social networks (just to mention a few elements) and one can see that the common citizen is unaware of how much of his/her will does not belong to him/her. This fact has a meaningful impact on our social, political, economic and private life. We are not doing enough to attend to this issue, and it is getting very late".

Llewellyn Kriel, CEO of TopEditor International, a media services company based in Johannesburg, South Africa, wrote, "Current developments do not augur well for the fair growth of AI. Vast swaths of the population simply do not have the intellectual capacity or level of sophistication to understand 1) the technology itself and 2) the implications of its safe use. This entrenches and widens the digital divide in places like Africa. The socio-political implications of this breed deep primitive superstition, racial hatred toward whites and Asians who are seen as techno-colonialists and the growth of kleptocracies amid the current mushrooming of corruption".

Steven Thompson, an author specializing in illuminating emerging issues and editor of "Androids, Cyborgs, and Robots in Contemporary Culture and Society," wrote, "The keyword from the query is 'dependence'. I published pioneering quantitative research on internet addiction and dependency in 1996, and followed up 15 years later with a related, updated research talk on the future of AI and internet dependency at a UNESCO-sponsored conference on information literacy in Morocco. My expertise is in ethical and technological issues related to moving the internet appliance into the human body. … The internet is moving into the human body, and, in that process, societal statuses are altered, privileging some while abandoning others in the name of emerging technologies, and the global order is restructuring to the same effect. Think of net neutrality issues gone wild, corporately and humanly sustained with the privileges such creation and maintenance affords some members of society. Now think of the liberty issues arising from those persons who are digital outcasts, and wish to not be on the grid, yet will be forced to do so by society and even government edicts".

Alan Mutter, a longtime Silicon Valley CEO, cable TV executive and now a teacher of media economics and entrepreneurism at the University of California, Berkeley, said, "The danger is that we will surrender thinking, exploring and experimentation to tools that hew to the rules but can't color outside the lines. Would you like computers to select the president or decide if you need hip surgery?"

Dan Geer, a respondent who provided no identifying details, commented, "If you believe, as do I, that having a purpose to one's life is all that enables both pride and happiness, then the question becomes whether AI will or will not diminish purpose. For the irreligious, AI will demolish purpose, yet if AI is truly intelligent, then AI will make serving it the masses' purpose. Ergo …"

Cristobal Young, an associate professor of sociology at Cornell University specializing in economic sociology and stratification, commented, "I mostly base my response [that tech will not leave most people better off than they are today] on Twitter and other online media, which were initially praised as 'liberation technology'. It is clear that the internet has devastated professional journalism, filled the public sphere with trash that no one believes and degraded civil discourse. This isn't about robots, but rather about how humans use the internet. Donald Trump himself says that without Twitter, he could never have been elected, and Twitter continues to be his platform for polarization, insult and attacks on the institutions of accountability".

David J. Krieger, co-director of the Institute for Communication & Leadership in Lucerne, Switzerland, wrote, "The affordances of digital technologies bind people into information networks such that the network becomes the actor and intelligence as well as agency are qualities of the network as a whole and not any individual actors, whether human or non-human. Networks will have access to much more information than do any present-day actors and therefore be able to navigate complex environments, e.g., self-driving cars, personal assistants, smart cities. Typically, we will consult and cooperate with networks in all areas, but the price will be that we have no such thing as privacy. Privacy is indeed dead, but in the place of personal privacy management there will be network publicy governance ['publicy' is the opposite of privacy]. To ensure the use of these technologies for good instead of evil it will be necessary to dismantle and replace current divides between government and governed, workers and capitalists as well as to establish a working global governance".

Wendy M. Grossman, author of "net.wars" and technology blogger, wrote, "2030 is 12 years from now. I believe human-machine AI collaboration will be successful in many areas, but that we will be seeing, like we are now over Facebook and other social media, serious questions about ownership and who benefits. It seems likely that the limits of what machines can do will be somewhat clearer than they are now, when we're awash in hype. We will know by then, for example, how successful self-driving cars are going to be, and the problems inherent in handing off control from humans to machines in a variety of areas will also have become clearer. The big fight is to keep people from relying on experimental systems and turning off the legacy ones too soon – which is our current situation with the internet".

Karl M. van Meter, founding editor of the Bulletin of Sociological Methodology and author of "Computational Social Science in the Age of Big Data," said, "The well-being of the world's population depends on governments making 'intelligent' decisions based on AI or other means. Moreover, environmental change may well be the determining factor for future well-being, with or without 'intelligent' decisions by world governments".

Andrew Whinston, computer science professor and director of the Center for Research in Electronic Commerce at the University of Texas at Austin, said, "There are several issues. First, security problems do not get the attention needed. Secondly, there may be use of the technology to control the population – as we see developing in China. AI methodology is focused on prediction, at least so far, so methods to improve health or general welfare are lacking. Deep learning, which is getting the big hype, does not have a clear foundation. That makes it scientifically weak".

An information administration manager responded, "We cede more and more decision-making and policy making to self-interested parties in the private sphere. Our institutions are insufficiently nimble to keep up with the policy questions that arise and attempts to regulate new industries are subverted by corrupt money politics at both the federal and state levels".

An internet pioneer said, "Nothing in our current social, economic or political structures points to a positive outcome. There is no evidence that more AI will improve the lives of most people. In fact, the opposite is likely to be the case. There will be more unemployment, less privacy, etc".

The following one-liners from anonymous respondents also tie into human agency:

• An Internet Hall of Fame member commented, "AI will not leave most people better off than they are today because individuals will not be able to control their lives".
• A professor of AI at a university in Italy said, "Development has brought humanity past the boundary, the survival limit; it is too easy to control technology in ways that are dangerous for people".
• An assistant professor of social justice wrote, "Technology magnifies what exists (for good or bad). There is simply more bad than good to be magnified".
• A professor of digital humanities at a Silicon-Valley-area university said, "Given increasing income disparity in much of the world, my fear is that AI will be used to repress the disenfranchised and create even more privilege for the few".
• A distinguished engineer and chief scientist at major technology companies commented, "Large actors will use AI for their benefit. Individual customers may have some benefits as a side effect, at a cost of lower autonomy".
• A professor of electrical engineering and innovation based in Europe said, "People will lose control of their lives, which will remain in the hands of a small group of experts or companies".
• A respondent based in Turkey wrote, "Due to unknown logic of algorithms we will lose our autonomy over our lives and everyday life decisions; humankind is depending on AI and not learning to be algorithmically literate".
• An engineer and chief operating officer said, "AI will be used to suppress rights".
• A technology fellow for a global organization commented, "I fear that AI will control many background choices with great implicating effects".

Other anonymous respondents commented:

• "More will be delegated to technology – smartphones, software. People will stop thinking or caring about 'control' and just delegate everything to 'the system'".
• "You can deploy most any technology in ways that enhance freedom [and] autonomy [or] have the opposite effect".
• "With China aiming to 'win' the AI lead, I have serious doubts that any benefits will outweigh the negative effects on human rights for a majority of people".
• "AI is not intelligent, it is human-made, and therefore biased and unreliable, it cannot do now what it is claimed it can do".
• "Provided we are still locked in capitalism I do not see how technology will help people stay engaged and empowered in our society".
• "My fear is that AI will be developed too quickly and that there may be severe repercussions once the genie is out of the bottle".