The sizeable majority of experts surveyed for this report envision major advances in robotics and artificial intelligence in the coming decade. In addition to asking them for their predictions about the job market of the future, we also asked them to weigh in on the following question:
To what degree will AI and robotics be parts of the ordinary landscape of the general population by 2025? Describe which parts of life will change the most as these tools advance and which parts of life will remain relatively unchanged.
These are the themes that emerged from their answers to this question.
AI and robotics will be integrated into nearly every aspect of most people’s daily lives
Jeff Jarvis, director of the Tow-Knight Center for Entrepreneurial Journalism at the City University of New York, wrote, “Think ‘Intel Inside’. By 2025, artificial intelligence will be built into the algorithmic architecture of countless functions of business and communication, increasing relevance, reducing noise, increasing efficiency, and reducing risk across everything from finding information to making transactions. If robot cars are not yet driving on their own, robotic and intelligent functions will be taking over more of the work of manufacturing and moving.”
Stowe Boyd, lead researcher for GigaOM Research, predicted, “Pizzas will not be delivered by teenagers hoping for a tip. Food will be raised by robotic vehicles, even in small plot urban farms that will become the norm, since so many people will have lost their jobs to ‘bots. Your X-rays will be reviewed by a battery of Watson-grade AIs, and humans will only be pulled in when the machines disagree. Robotic sex partners will be a commonplace, although the source of scorn and division, the way that critics today bemoan selfies as an indicator of all that’s wrong with the world.”
K.G. Schneider, a university librarian, wrote, “By 2025 AI, robotics, and ubiquitous computing will have snuck into parts of our lives without us understanding to what extent it has happened (much as I just went on a camping trip with a smartphone, laptop, and tablet).”
Lillie Coney, a legislative director specializing in technology policy in the U.S. House of Representatives, replied, “It is not the large things that will make AI acceptable it will be the small things—portable devices that can aid a person or organization in accomplishing desired outcomes well. AI embedded into everyday technology that proves to save time, energy, and stress that will push consumer demand for it.”
JP Rangaswami,chief scientist for Salesforce.com, wrote,“Traditional agriculture and manufacturing will both be affected quite heavily, with AI and robotics having greater roles to play at scale, while high-touch ‘retro’ boutiques will exist for both sectors. Service sector impact will continue to be lower in relative terms; knowledge/information worker sector impact, on the other hand, will be transformational.”
A Syracuse University professor and associate dean for research wrote, “Robots and AI are moving beyond simple rules into framed judgment spaces. There will be several spectacular failures (to give voice to the dystopian seers) and so many subtle impacts. I see them in public transport, long-distance driving, traffic routing, and car-to-car interactions. I also see them moving into the built environment through post-market sensor networks reflecting energy monitoring, maintenance for household appliances, and supporting more distributed education. My expectation is that much of medicine will be in the midst of a transformation based on better sensors tied to more powerful analytics.”
David Clark, a senior research scientist at MIT’s Computer Science and Artificial Intelligence Laboratory, noted that AI is already a part of daily life for many users: “AI methods and techniques are already part of the ordinary landscape. The problem with the term ‘AI’ is that it is constantly redefined to describe things we don’t yet know how to do well with computers. Things like speech recognition (like Siri), image recognition (face recognition in consumer cameras), and the like used to be hard AI problems. As they become practical commercial offerings, they spin off as their own disciplines.”
However, some experts sounded a note of concern that the gains from these new advances risk being limited only to those with the financial resources to afford the latest technologies, which may reinforce economic inequality.
The CEO of a professional not-for-profit society responded, “We will have more and more robots and AI in our lives, although I fear the benefits will accrue to the top 1-2% who can afford the gadgets.” And an information science professional and leader for a national association wrote, “In terms of day-to-day living, AI and robotics could easily be something that only the 1% can afford or have access to. In fields like medicine, though, advances have the potential to help everyone.”
A journalist, editor, and leader of an online news organization wrote, “Typically, this will depend on socioeconomics. The rich will spend almost no time doing things that can be automated; the poor will continue as is, more or less, although with superior communication abilities.”
These technologies will be integrated so completely as to be nearly invisible to most users most of the time
Depictions of robotics and artificial intelligence in popular culture often lean towards powerful anthropomorphic robots (Transformers, The Terminator) and hulking mainframes with human-like intelligence (HAL in 2001). But many of the experts who responded to this survey expect technology to evolve in the opposite direction, with machine intelligence being hidden deep in the complex workings of outwardly simple or even invisible devices and digital interactions.
John Markoff, senior writer for the Science section of the New York Times, likens this process to a kind of magic: “Over the next decade the ubiquitous computing era will come into full force. Computers will ‘disappear’ and ordinary objects will become magic. Significantly, Steve Jobs was the first one to really understand this. But the pace is relentless.”
Nishant Shah, a visiting professor at The Centre for Digital Cultures at Leuphana University in Germany, wrote, “The primary function of care robots or companion AIs is to be invisible. They are already ubiquitous in the world that we live in, but largely they work under the surface, and below the networks. Advancements in nanotechnologies and wearable computing are going more in the direction of creating tools that we do not see.”
David Organ, CEO of Dotsub, wrote, “The progressive availability of more and more robust AI systems, with deeper predictive power and broader contextual understanding will make them almost invisible. The people who are not specialists of the field will react to their advances being pointed out with a sense of natural acceptance because the progressive arrival of better and better features and performance will have created a sense of familiarity. It will be natural to talk to computers of any shape, and expect them to understand the words, and the meaning, and to establish a dialog leading rapidly to the desired goal.”
Elizabeth Albrycht, a senior lecturer in marketing and communications at the Paris School of Business, replied, “By 2025 we may well be witnessing the disappearance of AI and robotics into the ordinary landscape as they follow the usual path of technology. First we see it, then it becomes invisible as it integrates into the landscape itself.”
The fact that the “invisible” technologies of the future may be doing jobs currently held by human beings was not lost on some respondents:
Olivier Crepin-Leblond, managing director of Global Information Highway Ltd. in London, UK, offered similar thoughts when he predicted that, “…An enormous amount of automata will replace humans—from automated passport gates at border control, to onsite vending machines, automated floor cleaners, window cleaning machines, driving trains, cars etc. Our day-to-day life will remain the same, but those jobs performed in the past by what some call ‘invisible people,’ will be performed by ‘invisible robots.’ How many people remember the face of the ticket collector on their train? That’s what I mean by ‘invisible people.’ Now the life of the people performing the work of ‘invisible people’ will be heavily affected as they’ll be out of work. The life of others too: I rely on these ‘invisible people’ to bring a human face to the world and to my life—a hello, a smile, a thanks.”
Driving, transportation, and logistics will experience dramatic changes
Many respondents predicted that self-driving cars will enter the ordinary landscape in a meaningful way within the next decade. Howard Rheingold, a pioneering Internet sociologist and self-employed writer, consultant, and educator, expressed his belief that this can only be for the best: “I, for one, welcome our self-driving automobile overlords. How could they possibly do a worse job than the selfish, drugged, drunk, and distracted humans who have turned our roads into bloodbaths for decades?” A self-employed programmer and Web developer offered similar thoughts: “We might wonder how we accepted so many car accidents in 2013 and wonder why we even bothered to perform the necessary but menial task of, say, parking.”
Other respondents imagined a future with many more driverless cars, but many fewer truck drivers, delivery people, and taxi operators. danah boyd, a research scientist for Microsoft, responded, “There will be a lot more automation but much of it will be as invisible as it is now. So in that sense, yes, it will be part of the ordinary landscape. The biggest change will be to the movement of atoms—food, consumer goods, etc. The majority of the disruption will be at the blue-collar level, and I suspect that the biggest impact will be in warehouses (or ‘fulfillment centers’).”
Tom Standage, digital editor for The Economist, wrote, “Self-driving vehicles promise to upend existing approaches to car ownership, car design, car sales and insurance, urban planning, logistics, deliveries, taxi services, etc. That will be a big change, as significant as the advent of smartphones.” And Linda Rogers, the founder of Music Island in Second Life and grant writer for Arts for Children and Youth in Toronto, concurred: “We already see it in grocery scanners, bank machines and can extrapolate from there as automated parking lots add robotic valet service, subway lines no longer require drivers, and garbage pickup services are robot-controlled.”
Other respondents envisioned a wide range of impacts that might arise from the driverless car revolution—from the economic to the cultural.
Andrew Rens, chief council at the Shuttleworth Foundation, wrote, “AI and robotics will change the way that Western society thinks about cars. Once control over driving passes to software the romance of cars will diminish. There will be far less cachet in owning large and powerful cars since the riding (rather than driving) experience will be indistinguishable.”
Robert Bell of IntelligentCommunity.org responded, “Technology will continue to make things better, faster, cheaper and safer: the impact of self-driving cars alone will be immense in terms of reduced traffic congestion, lower costs for insurance and transport, and driver safety.”
A professor at Aoyama Gakuin University in Tokyo, Japan, wrote, “Self-driving cars may change a lot. Car renting and sharing will be far easier and thus more popular, which will be a good thing. On the other hand, spending long hours in cars will be easier (because you can sleep or work or watch a movie while driving), which is not necessarily a good thing.”
Intelligent agents will increasingly manage our day-to-day lives and be omnipresent in our homes
As computer intelligence becomes increasingly integrated in daily life, a number of experts expect major changes in the way people manage their households and day-to-day lives.
Hal Varian, chief economist for Google, views the current wave of smartphone-enabled assistants as the tip of the iceberg: “We will rely on personal assistance from devices such as Google Now, Siri, Watson, etc. Much of the interaction will be verbal, so this will look a lot like the Star Trek computer interaction. We will expect computers that we meet to know us and our history of interaction with them. In general, they will infer what we want, and our role is simply to refine and verify that expectation. We will be well on our way to universal access to all human knowledge via the worldwide network of mobile devices and data centers. Day-to-day interaction with devices and data will be by voice. One industry that will be hugely affected is education: what should be people be taught when they can access all human knowledge all the time?”
A CEO for a company that builds intelligent machines wrote, “The creative class by 2025 will have a digital assistant in their work and personal lives who all but replaces what we think of today as administrative help. That entity (actually a collection of distributed software) will answer phones, schedule appointments (handling the logistics far more accurately than any human), manage the care and maintenance of that person’s living quarters and work environment, do the shopping and (where appropriate) be responsible for managing that person’s financial life.”
Frederic Litto, a professor emeritus at the University of Sao Paulo in Brazil, responded, “It will probably be in ‘concierge’-type services—that is, everyone’s device (be it smartphone, tablet, or Dick-Tracy-on-the-wrist devices) will have built-in applications to remind users of things to be done, and featuring unlimited lists of contacts, past and present, as well as the contents of global and local reference works, model decision-trees, and other handy information devices. Concierge-type services will give citizens greater autonomy in everyday activities, as well as in highly specialized professional activities (like on-the-wrist ‘specialist systems’ for medical diagnoses).”
A technology policy expert wrote, “Where I think the public will see it more is via mobile devices and home automation. I expect that new construction will include learning thermostats, embedded smoke detectors, smart appliances, automated door locks, etc., all run by apps.”
A general manager for Microsoft replied, “Robotics and AI will have a broader role in daily life. We are already seeing trends in home automation and maintenance, for example, that if extrapolated to 2025 at the same development rate will create substantially different experiences in a future-modern home.”
Large swathes of the service sector—both online and off—will be impacted
Many experts anticipate that advances in AI and robotics will produce dramatic changes in the service industry by 2025. Glenn Edens, a director of research in networking, security, and distributed systems within the Computer Science Laboratory at PARC, a Xerox Company, predicted, “It is likely most consumer services (banking, food, retail, etc.) will move to more and more self-service delivery via automated systems.”
Joe Touch, director of the Information Sciences Institute’s Postel Center at the University of Southern California, replied, “They will continue to replace certain simple tasks, including, I would expect, mail and package delivery, and will increasingly shift from warehouses to public shopping areas (e.g., restocking shelves, or avoiding the need for bulk shelf displays in stores altogether). Interfaces will increasingly involve speech recognition and vision, interacting with people on more ‘human’ terms.”
An anonymous respondent wrote, “A large portion of service jobs may be taken over by AI—ticket clerks at movie theaters, bank tellers, automated clerks in most service positions. Once we begin to program the software to manage intelligent response to human interaction we may find that simpler tasks may be taken over completely by AI.”
Per Ola Kristensson, a lecturer in human-computer interaction at the University of St Andrews, UK, responded, “While automation will be less than perfect by 2025, we are likely to witness a trend in which routine white-collar jobs, such as routine legal work, accounting, and administration, will be replaced by AI tools.”
Several experts predicted that most of our online and telephone interactions with customer service “personnel” in the future will be with intelligent algorithms. Judith Donath, a fellow at Harvard University’s Berkman Center for Internet & Society, responded, “Conversations with intelligent-seeming agents will be far more common. It will be frequently difficult to tell (online at least) if you are speaking/chatting with a person or program—and people will have become accustomed to this and will have ceased to care in many cases. Dealing with a machine will often be more efficient, and many people will come to use the sort of shorthand commands—no greetings or niceties, imperative forms—that they use with AI agents with anyone in a subordinate position.”
Thomas Haigh, an information technology historian and associate professor of information studies at the University of Wisconsin, observed, “AI will make it increasingly easy to interface with computer systems in flexible ways. Automated decision making has largely automated fairly complex business processes like credit card applications, and coupled with big data will continue to displace human judgment in routine transactions.”
Mary Joyce, an Internet researcher and digital activism consultant, replied, “Customer service, which firms have been trying to automate and outsource to the frustration of customers, are likely to adopt interactive automated customer service agents more sophisticated than current voice-recognition systems.”
Advances in AI and robotics will be a boon for the elderly, disabled, and sick
A number of experts surveyed predicted that caring for the sick, elderly, and physically challenged will be revolutionized by advances in robotics.
David Clark, a senior research scientist at MIT’s Computer Science and Artificial Intelligence Laboratory, said, “I like to consider questions such as this by asking what problem needs a solution. I believe that one reason the ‘smart home’ has not taken off so well is that the dumb house is good enough. I think commuting is a problem (so self-driving cars as well as telework will be popular). We will see robots in health care and care of the elderly. But these may not be humanoid robots, but devices designed to work in specialized spaces designed for them.”
Jonathan Grudin, principal researcher for Microsoft, responded, “I expect more robotic assistance for the elderly and infirm, because the demands are manageable and the need is increasing.”
Gary Kreps, professor of communication and director of the Center for Health and Risk Communication at George Mason University, wrote, “Smart interactive virtual human agents will be a common part of modern life, providing the public with access to relevant information and support wherever they are and whenever they need help. This will be particularly important in providing consumers with access to relevant health information and support for making important health promotion decisions and guiding self-care and care for loved ones at home. This will improve the quality of self- and other-care, as well as enhance adherence with health regimens in the future.”
The head of the department of communication at a top U.S. university wrote, “The low-hanging fruit for AI and robotics are areas of labor that still involve high degrees of routine. Special areas of need in the developed world involve domestic assistance for aging populations and other vulnerable groups.”
Daren C. Brabham, assistant professor at the Annenberg School for Communication & Journalism, University of Southern California, says these developments could help expand health coverage in hard-to-reach populations: “I see robotics/AI taking a stronger hold in medicine, both in medical research and testing and in doctor-patient interactions. On this latter point, basic telemedicine applications/robots will serve a significant portion of healthcare needs for rural and poor populations by 2025, with robot-doc-in-a-box pods dispersed throughout the country that can automatically take blood pressure, draw blood, and other simple diagnostic procedures.”
The CEO of a software technology company and active participant in Internet standards development, responded, “Hopefully one of the areas where this will have most impact is the medical field—this is an area where there are high costs, a shortage of highly skilled people and a growing demand for advanced and complex services.”
Janet Kornblum, a self-employed media trainer and journalist, observed, “Robotics is already a part of our landscape. Medically, many of us will use intelligent devices that help us function, be they smart replacements for little reminders that tell us when to take our pills, etc. Will robots be caring for us? Maybe. I think medical robots will lead the way.”
Larry Magid, a technology journalist and an Internet safety advocate, responded, “People won’t have to drive cars unless they want to, and senior citizens and people with disabilities will be able to live more independently.”
Minority viewpoint: expect these changes to be gradual and incremental
Although most of our respondents expect dramatic advances in AI and robotics in the coming decade, some expect that these changes will occur much more gradually.
A program director focusing on ICT standards policy, Internet Governance and other issues wrote, “It will still be limited. Although we can already do some pretty cool stuff, there will still be plenty of kinks and bugs and vulnerabilities that need to be resolved before market confidence will be widespread.”
The former chair of an IETF working group wrote, “Change will continue to be pretty gradual in the next 12 years. AI and robotics are making great strides but will not suddenly take over a lot of domestic / household functions. The areas that border on factory automation are the candidates for change—perhaps low skill assembly and clothing fabrication jobs will be affected next.”
The principal architect at an enterprise computing firm wrote, “Partial solutions will dominate—parking aids, automatic mowers and the like. People underestimate the value and convenience of cheap labor.”
An executive at an Internet top-level domain name operator replied, “Considering the percentage of the U.S. population that remains offline to this day, we shouldn’t get ahead of ourselves in predicting some kind of I, Robot-type world within 10 years—which is less time than the commercial Internet has been available.”
“Out of the box” responses
Along with the major themes highlighted above, several experts made thought-provoking predictions about the future of AI and robotics that did not fit cleanly into any of these major categories. Among the more interesting ideas proposed:
Warfare and police work will be increasingly mechanized
Several experts expressed concerns about increased mechanization of warfare, surveillance, and police work. Marc Rotenberg, president of the Electronic Privacy Information Center (EPIC), noted, “You will see early versions of RoboCop on city streets. Looking at the current evolution of surveillance drones we can anticipate that that they will have the ability to interpret sound and images. They will also sense chemical compositions to help identify explosive and other harmful elements. They will likely have both infrared detection as well as the ability to see through solid materials and detect heat signatures. They will certainly have facial recognition capabilities and be integrated with a national biometric center. An interesting question is whether they will also have non-lethal weapons, such as tasers. Several incidents of attacks on robots will be reported.”
[to the increased use of AI and robotics]
Increased automation will spark a “machine free” movement
Andrew Rens, chief council at the Shuttleworth Foundation, wrote, “The rise of AI and robots will also likely change extreme sports and outdoor pursuits not by increased reliance on AI and robotics but by provoking a movement to purge extreme sports of them. Extreme sports and outdoor pursuits such as hunting are one area of life that encourages immersion in the natural world, self-reliance, and human excellence. As other areas of life become increasingly dominated by machines that are faster, more accurate, and more reliable than humans, outdoor pursuits and extreme sports will become increasingly valuable to a substantial minority as they seek to carve out space from a frenetically connected world. The perception of extreme sports and outdoor pursuits as a machine-free zone will provoke debate about the ethics of relying on machines. A significant minority of sportspeople will attempt complete human self-reliance, even refusing current technologies such as GPS except in emergencies.”
The nature of memory and imagination will change
Tiffany Shlain, filmmaker, host of the AOL series The Future Starts Here, and founder of The Webby Awards, responded, “The parts of life that will change most will be our sense of memory and interactions with new ideas. We will have robotic aids to help us remember facts, memories and access to ideas that will give our minds amplified abilities. Everyone human on the planet who wants to be connected will definitely be connected by 2025. This intersection and recall of diverse ideas will have led to great innovation. What will not change—is our human desire for authentic connection and eye contact.”
The world will contain more “magic”
A research scientist working at a major search engine company responded, “There will be more ‘magic’ in the world. I mean this in the sense that more actions will be taken for us, to us, by our systems that will not have explanations attached or perceivable reasons why they’re being taken. Example: recommender systems will become everyday interactions multiple times per day. In many cases, even the software engineers have no idea, really, why a particular recommendation is being made. That’s surprising, and magical. You decide if it’s net good or not. Opinions will be split.”
How will we interact with each other?
Vytautas Butrimas, the chief adviser to a major government’s ministry, expects technological advances to increase the distance between people: “AI and robotics will change the way we interact with other members of society. The tendency will be toward more social isolation and fewer human-to-human contacts taking place. Just look at what is happening at our airport waiting lounges. People sit next to each other but the interaction is not taking place with the neighbor sitting nearby but with a device communicating to some other device. The world will be more bureaucratic and ‘cold’ in 2025 than it is today.”
On the other hand, a self-employed writer, researcher, and consultant wrote, “I expect there will need to be a ‘human please’ option on most commercial and transport interfaces as there are always unpredictable elements. I also expect that consumers will demand tech-free places like restaurants where they can interact with each other.”
As more daily activities are automated, human interaction may become a luxury
Judith Donath, a fellow at Harvard University’s Berkman Center for Internet & Society, responded, “Even simple technologies have been doing this—most of what was a secretary’s job has been replaced by answering machines and Word. Robots will be able to stock store shelves and check out and bag groceries and other store purchases. They’ll do much of today’s custodial work, delivery services, and transportation. Customer service will be almost entirely done with scripted agents. Software agents will work their way up from the crowd scenes in movies to smaller speaking roles, and eventually to fully automated ‘live’ films. Employment will be mostly very skilled labor—and even those will jobs will be continuously whittled away by increasingly sophisticated machines. Live, human salespeople, nurses, doctors, actors will be symbols of luxury, the silk of human interaction as opposed to the polyester of simulated human contact.”
Andrew Bridges, a partner and Internet law litigator and policy analyst at Fenwick & West LLP, wrote, “’Brain work’ will increasingly become a commodity as computing power enables more artificial intelligence. We already see Google Translate displacing translators, investment advice algorithms displacing investment advisors, automated landing systems replacing airplane piloting skills, and so forth. I expect that the world will become increasingly divided between ‘standard’ service and ‘concierge’ service in many aspects, with standard service left entirely to the machines and concierge service resting more upon human relationships.”
On the political economy of self-driving cars
Doc Searls, director of ProjectVRM at Harvard University’s Berkman Center for Internet & Society, wrote at length on what might need to happen for self-driving cars to reach wide-scale adoption:
“First, prototypes of self-driving cars are already here, thanks to Google and a few others. Driving, however, has been a human activity from the start, with century-old norms and a regulatory framework spanning global, national, interstate, state and local jurisdictions. Getting self-driving cars to work within all of that, and for regulations to adapt as well, seems a tall order that will require a lot of time and many trials and errors along the way. If it happens, 2025 is probably too early a date for seeing lots of self-driving cars, except perhaps in a few isolated geographies.
So far, the strongest arguments for self-driving cars are savings and safety. So let’s say savings is covered, and the cost of owning a self-driving car is cheaper than owning a conventional one. On the safety side, according the June 2012 NHTSA Traffic Safety Facts report (pdf), 5,419,000 crashes in the U.S. killed 32,885 people and injured 2,239,000. That should make a good argument (especially around the number of crashes); but those numbers—fatalities especially—have been going down. And how much do people, or industry, actually care?
Consider that the third leading cause of death in the U.S., according to the Center for Disease Control (after heart attacks and cancer) is medical error. The Journal of Patient Safety puts the medical error fatality total in the U.S. at between 210,000 and 440,000 per year. Many studies suggest that a large percentage of those deaths could be prevented with better patient information, which today is scattered among many health care providers with incompatible systems that barely communicate with each other, much less doctors and patients. Yet reform, both within the health care industry and within legislative and regulatory systems, has ranged from difficult (e.g. the Affordable Care Act) to impossible—at least in the U.S.
Would the insurance industry, which basically runs health care in the U.S. (the vast majority of payments are B2B, not C2B), welcome self-driving cars? If so, that would help the cause. But insurance itself is a shell game, depending on a high degree of knowledge asymmetry to the advantage of insurance companies over other players, including car makers and owners. Insurance companies might prefer the game they know over one that would put far more power in other hands, such as the government, car owners and the likes of Google. If that turns out to be the case, the car insurance industry might be as reluctant to reform driving as the health insurance industry has been to reform medical care, with far more cause.”
