Artificial Intelligence and the Future of Humans

One of the chief fears about today's technological change is the possibility that autonomous hardware and software systems will cause millions of people globally to lose their jobs and, as a result, their means for affording life's necessities and participating in society. Many of these experts say new jobs will emerge along with the growth of AI just as they have historically during nearly every human transition to new tools.

Brad Templeton, chair for computing at Singularity University, said, "While obviously there will be good and bad, the broad history of automation technologies is positive, even when it comes to jobs. There is more employment today than ever in history".

Ben Shneiderman, distinguished professor and founder of the Human Computer Interaction Lab at the University of Maryland, said, "Automation is largely a positive force, which increases productivity, lowers costs and raises living standards. Automation expands the demand for services, thereby raising employment, which is what has happened at Amazon and FedEx. My position is contrary to those who believe that robots and artificial intelligence will lead to widespread unemployment. Over time I think AI/machine learning strategies will become merely tools embedded in ever-more-complex technologies for which human control and responsibility will become clearer".

Robert D. Atkinson, president of the Information Technology and Innovation Foundation, wrote about how advances in AI are essential to expanded job opportunities: "The developed world faces an unprecedented productivity slowdown that promises to limit advances in living standards. AI has the potential to play an important role in boosting productivity and living standards".

Toby Walsh, a professor of AI at the University of New South Wales in Australia and president of the AI Access Foundation, said, "I'm pessimistic in short term – we're seeing already technologies like AI being used to make life worse for many – but I'm optimistic in long term that we'll work out how to get machines to do the dirty, dull, dangerous and difficult, and leave us free to focus on all the more-important and human parts of our lives".

Yet many others disagree. Some fear the collapse of the middle class and social and economic upheaval if most of the world's economic power is held by a handful of technology behemoths that are reaping the great share of financial rewards in the digital age while employing far fewer people than the leading companies of the industrial age. A fairly large share of these experts warn that if steps are not taken now to adjust to this potential future that AI's radical reduction in human work will be devastating".

David Cake, an leader with Electronic Frontiers Australia and vice-chair of the ICANN Generic Names Supporting Organization Council, wrote, "The greatest fear is that the social disruption due to changing employment patterns will be handled poorly and lead to widespread social issues".

Jerry Michalski, founder of the Relationship Economy eXpedition, said, "We're far from tipping into a better social contract. In a more-just world, AI could bring about utopias. However, many forces are shoving us in the opposite direction. 1) Businesses are doing all they can to eliminate full-time employees, who get sick and cranky, need retirement accounts and raises, while software gets better and cheaper. The precariat will grow. 2) Software is like a flesh-eating bacterium: Tasks it eats vanish from the employment landscape. Unlike previous technological jumps, this one unemploys people more quickly than we can retrain and reemploy them. 3) Our safety net is terrible and our beliefs about human motivations suck. 4) Consumerism still drives desires and expectations".

James Hendler, professor of computer, web and cognitive sciences and director of the Rensselaer Polytechnic Institute for Data Exploration and Application, wrote, "I believe 2030 will be a point in the middle of a turbulent time when AI is improving services for many people, but it will also be a time of great change in society based on changes in work patterns that are caused, to a great degree, by AI. On the one hand, for example, doctors will have access to information that is currently hard for them to retrieve rapidly, resulting in better medical care for those who have coverage, and indeed in some countries the first point of contact in a medical situation may be an AI, which will help with early diagnoses/prescriptions. On the other hand, over the course of a couple of generations, starting in the not-too-distant future we will see major shifts in work force with not just blue-collar jobs, but also many white-collar jobs lost. Many of these will not be people 'replaced' by AIs, but rather the result of a smaller number of people being able to accomplish the same amount of work – for example in professions such as law clerks, physicians assistants and many other currently skilled positions we would project a need for less people (even as demand grows)".

Betsy Williams, a researcher at the Center for Digital Society and Data Studies at the University of Arizona, wrote, "AI's benefits will be unequally distributed across society. Few will reap meaningful benefits. Large entities will use AI to deliver marginal improvements in service to their clients, at the cost of requiring more data and risking errors. Employment trends from computerization will continue. AI will threaten medium-skill jobs. Instead of relying on human expertise and context knowledge, many tasks will be handled directly by clients using AI interfaces or by lower-skilled people in service jobs, boosted by AI. AI will harm some consumers. For instance, rich consumers will benefit from self-driving cars, while others must pay to retrofit existing cars to become more visible to the AI. Through legal maneuvering, self-driving car companies will avoid many insurance costs and risks, shifting them to human drivers, pedestrians and bicyclists. In education, creating high quality automated instruction requires expertise and money. Research on American K-12 classrooms suggests that typical computer-aided instruction yields better test scores than instruction by the worst teachers. By 2030, most AI used in education will be of middling quality (for some, their best alternative). The children of the rich and powerful will not have AI used on them at school; instead, they will be taught to use it. For AI to significantly benefit the majority, it must be deployed in emergency health care (where quicker lab work, reviews of medical histories or potential diagnoses can save lives) or in aid work (say, to coordinate shipping of expiring food or medicines from donors to recipients in need)".

Nathaniel Borenstein, chief scientist at Mimecast, wrote, "Social analyses of IT [information technology] trends have consistently wildly exaggerated the human benefits of that technology, and underestimated the negative effects. … I foresee a world in which IT and so-called AI produce an ever-increasing set of minor benefits, while simultaneously eroding human agency and privacy and supporting authoritarian forms of governance. I also see the potential for a much worse outcome in which the productivity gains produced by technology accrue almost entirely to a few, widening the gap between the rich and poor while failing to address the social ills related to privacy. But if we can find a way to ensure that these benefits are shared equally among the population, it might yet prove to be the case that the overall effect of the technology is beneficial to humanity. This will only happen, however, if we manage to limit the role of the rich in determining how the fruits of increased productivity will be allocated".

Andrea Romaoli Garcia, an international lawyer active in internet governance discussions, commented, "AI will improve the way people make decisions in all industries because it allows instant access to a multitude of information. People will require training for this future – educational and technological development. … This is a very high level of human development that poor countries don't have access to. Without proper education and policies, they will not have access to wealth. The result may be a multitude of hungry and desperate people. This may be motivation for wars or invasion of borders. Future human-machine interaction (AI) will only be positive if richer countries develop policies to help poorer countries to develop and gain access to work and wealth".

Josh Calder, a partner at the Foresight Alliance, commented, "The biggest danger is that workers are displaced on a mass scale, especially in emerging markets".

Jeff Johnson, computer science professor at the University of San Francisco, previously with Xerox, HP Labs and Sun Microsystems, responded, "I believe advances in AI will leave many more people without jobs, which will increase the socioeconomic differences in society, but other factors could help mitigate this, e.g., adoption of guaranteed income".

Alan Bundy, a professor of automated reasoning at the University of Edinburgh, wrote, "Unskilled people will suffer because there will be little employment for them. This may create disruption to society, some of which we have already seen with Trump, Brexit, etc".

Peter Levine, associate dean for research and professor of citizenship and public affairs in Tufts University's Tisch College of Civic Life, wrote, "Being a fully-fledged citizen has traditionally depended on work. I'm worried that rising levels of non-employment will detract from civic engagement. Also, AI is politically powerful and empowers the people and governments that own it. Thus, it may increase inequality and enhance authoritarianism".

Hassaan Idrees, an electrical engineer and Fulbright Scholar active in creating energy systems for global good, commented, "I believe human-machine interaction will be more of [a] utility, and less fanciful as science fiction puts it. People will not need to see their physicians in person, their automated doctors making this irrelevant. Similarly, routine workplace activities like data processing and financial number crunching would be performed by AI. Humans with higher levels of intellect can survive this age, and those on the lower ends of spectrum of mental acumen would be rendered unnecessary".

Ethem Alpaydin, a professor of computer engineering at Bogazici University in Istanbul, responded, "As with other technologies, I imagine AI will favor the developed countries that actually develop these technologies. … For the developing countries, however, whose labor force is mostly unskilled and whose exports are largely low-tech, AI implies higher unemployment, lower income and more social unrest. The aim of AI in such countries should be to add skill to the labor force rather than supplant them".

Sam Ladner, a former UX researcher for Amazon and Microsoft, now an adjunct professor at Ontario College of Art and Design, wrote, "Technology is not a neutral tool, but one that has our existing challenges imprinted onto it. Inequality is high and growing. Too many companies deny their employees a chance to work with dignity, whether it be through providing them meaningful things to do, or with the basic means to live. AI will be placed on top of that existing structure. Those who already have dignified work with a basic income will see that enhanced; those who are routinely infantilized or denied basic rights will see that amplified. Some may slip into that latter category because their work is more easily replaced by AI and machine learning".

Jonathan Swerdloff, consultant and data systems specialist for Driven Inc., wrote, "The more reliant on AI we become, the more we are at the mercy of its developers. While AI has the ability to augment professionals and to make decisions, I have three concerns which make me believe it will not leave us better off by 2030. This does not address fears that anything run via AI could be hacked and changed by bad faith third parties. 1) Until any sort of self-policed AI sentience is achieved, it will suffer from a significant GIGO [garbage-in, garbage-out] problem. As AI as currently conceived only knows what it is taught, the seed sets for teaching must be thought out in detail before the tools are deployed. Based on the experience with Microsoft's Tay and some responses I've heard from the Sophia robot, I am concerned that AI will magnify humanities flaws. 2) Disparate access. Unless the cost for developing AI drops precipitously – and it may, since one AI tool could be leveraged into building further less expensive AI tools – access to whatever advantages the tools will bring will likely be clustered among a few beneficiaries. I view this akin to high frequency trading on Wall Street. Those who can, do. Those who can't, lose. 3) Tool of control. If AI is deployed to make civic or corporate decisions, those who control the algorithms control everything. In the U.S. we've recently seen Immigration and Customs Enforcement change its bond algorithm to always detain in every case".

Stuart A. Umpleby, a professor and director of the research program in social and organizational learning at George Washington University, wrote, "People who use AI and the internet will have their lives enhanced by these technologies. People who do not use them will be increasingly disconnected from opportunities. As the digital world becomes more complicated and remote from real-world experiences, the need will grow for people and software to make connections. There will be a need for methods to distinguish the real world from the scam world".

Simeon Yates, director of the Centre for Digital Humanities and Social Science at the University of Liverpool, said, "AI will simply increase existing inequalities – it, like the internet, will fail in its emancipatory promise".

Panagiotis T. Metaxas, author of "Technology, Propaganda and the Limits of Human Intellect" and professor of computer science at Wellesley College, responded, "There will be a lot of wealth that AI-supported devices will be producing. The new technologies will make it easier and cheaper to produce food and entertainment massively ('bread and circus'). This wealth will not be distributed evenly, increasing the financial gap between the top small percentage of people and the rest. Even though this wealth will not be distributed evenly, the (relatively small) share given to the vast majority of people will be enough to improve their (2018) condition. In this respect, the majority of people will be 'better off' than they are today. They may not feel better off if they were aware of the inequalities compared to the top beneficiaries, but they will not be aware of them due to controlled propaganda. Unfortunately, there will not be much they could do about the increased inequalities. Technologies of police enforcement by robots and lack of private communication will make it impossible for them to organize, complain or push for change. They will not be valued as workers, citizens or soldiers. The desire for democracy as we know it today will be coming to an end. Many will feel depressed, but medical products will make it easy for them to increase pleasure and decrease pain".

Grace Mutung'u, co-leader of the Kenya ICT Action Network, responded, "New technologies will more likely increase current inequalities unless there is a shift in world economics. From the experience of the UN work on Millennium Development Goals, while there has been improvement with the quality of life generally, low- and middle-income countries still suffer disparate inequalities. This will likely lead to governance problems. In any case, governments in these countries are investing heavily in surveillance which will likely have more negative effects on society".

Danny Gillane, a netizen from Lafayette, Louisiana, commented, "Technology promises so much but delivers so little. Facebook gave us the ability to stay in touch with everyone but sacrificed its integrity and our personal information in pursuit of the dollar. The promise that our medical records would be digitized and more easily shared and drive costs down still has not materialized on a global scale. The chief drivers of AI innovation and application will be for-profit companies who have shown that their altruism only extends to their bottom lines. Like most innovations, I expect AI to leave our poor even poorer and our rich even richer, increasing the numbers of the former while consolidating power and wealth in an ever-shrinking group of currently rich people".

A professional working on the setting of web standards wrote, "Looking ahead 12 years from now, I expect that AI will be enhancing the quality of life for some parts of some populations, and in some situations, while worsening the quality of life for others. AI will still be uneven in quality, and unevenly available throughout different parts of society. Privacy and security protections will be inadequate; data bias will still be common; many technologies and response patterns will be normed to the needs of the 'common denominator' user and misidentify or misinterpret interactions with people with disabilities or, if appropriately identifying their disability, will expose that information without user consent or control".

So many people included comments and concerns about the future of jobs for humans in their wide-ranging responses to this canvassing that a later section of this report has more expert opinions on this topic.

The following one-liners from anonymous respondents also tie into AI and jobs:

• An associate professor of computer science commented, "Machines will be able to do more-advanced work and improve accuracy, but this likely will expand manipulation of consumers/voters and automation may reduce available jobs".
• A director for a global digital rights organization said, "My concern is that human-machine collaboration will leave some of us far better off by automating our jobs, giving us more free and creative time, while doing little to improve the lives of billions of others".
• A professor expert in cultural geography and American studies said, "Given the majority human assumption that capitalism is something worth reproducing, the evacuation of most labor positions by AI would create vast poverty and cruelty by the ruling class".
• A lecturer in media studies based in New Zealand wrote, "The automation of large volumes of work by machine learning-based systems is unlikely to lead to an increase in social equity within a capitalist economy".
• A senior partner at one of the world's foremost management consulting firms commented, "AI will benefit businesses, the economy and people as consumers, but likely increase income/wage polarization so most people as workers may not benefit".
• An engineer and chief operating officer for project automating code said, "Those with the most money will leverage their position of power through AI; it will lead to possibly cataclysmic wealth disparity.
• A digital anthropologist for a major global technology company wrote, "The gap between those who benefit from advances in technology and those who do not have widened over the past three decades; I can't see an easy or quick reversal".

Other anonymous respondents commented:

• "Some will benefit, while others will suffer. The bifurcated economy will continue to grow. … Those at the bottom of the ladder will see greater numbers of jobs being taken away by technology".
• "All in all, AI can be of great use, but we need to be vigilant of the repercussions instead of constantly leaping 'forward' only to find out later about all of the negatives".
• "In the U.S., the blue-collar job wages have been stagnant since the 1970s despite all of the advances with the internet and mobile devices, so I am not optimistic regarding AI".
• "Wealth distribution will continue to widen as the rich get richer".
• "AI is going to lead to the destruction of entire rungs of the economy, and the best way to boost and economy while holding together a fractured economy is war".
• "Many people will no longer be useful in the labor market. Such rapid economic and social change will leave many frightened and angry".
• "In 12 years AI may be more disruptive than enabling, leaving many without work until they retrain and transition".
• "There could be a thinning out of the middle – middle management and class".
• "AI will increasingly allow low-quality but passable substitutes for previously-skilled labor".
• "There are significant implications for unskilled or easily-automated tasks on one end of the spectrum and certain types of analysis on the other that will be automated away. My concern is that we have no plan for these people as these jobs disappear".