Information is the most valuable thing there is. It’s why companies hire consultants just to share what they know, it’s how stock brokers and financial advisers are able to make a living without actually creating anything, and it’s why the big data and predictive analytics industry is projected to reach $50 billion by 2017. Most important of all, it’s how Biff Tannen was able to amass a fortune through gambling in Back to the Future Part II after borrowing the time machine and giving a younger version of himself a sports almanac from the future.
Big data is popular because it presents an opportunity to make the sports almanac from Back to the Future Part II real. Getting reliable information about the future is an especially enticing proposition because it’s something that people naturally are bad at. Biological evolution favored those who were able to recognize patterns that would lead their tribe to food or accurately predict what would happen if they tickled a bear, but the infinitely branching nature of consequence makes more sophisticated predictions almost impossible. Anyone who’s ever put money on a sporting event knows that outcomes, even binary ones, are only obvious after they’ve happened, no matter how sure you were that the Steelers would cover the spread.
Human pragmatism is planted firmly in the present, which also makes people bad at re-imagining context. Finding tonight’s dinner has taken precedence over almost everything else for thousands of years. When asked what the problems of the future will be, people will nearly always describe some variation of today’s problems. The challenge of modern man really lies in identifying what might be just around the corner.
What new problems will government face and what new jobs will be created to address them? The crystal ball in the basement office of Public CIO headquarters revealed how the coming decades will transform the public-sector IT workforce.
As valuable as information is, the value of a discovery is diminished almost to zero if there’s no follow-up action. Information is only valuable if it’s acted on. Predictive data engineers will ensure their organizations don’t squander data opportunities. Through this office, agencies will inspire behavioral change informed by data through a technique known as gamification.
Today’s data analytics platforms are like a beautiful car dashboard that displays measurements with extreme precision, but when the driver turns the steering wheel, there’s a six-month delay before anything happens, said Rajat Paharia, founder of gamification company Bunchball.
The talent of the predictive data engineer won’t just be in reading and analyzing data, Paharia said, but also in knowing how to work the buttons and levers of the organization to make change happen. Gamification, Paharia said, is the steering wheel that will allow an organization to change course immediately.
“What gamification does is it motivates behavior change using data as the raw material,” he explained. “It shouldn’t just be about looking at the data. It’s about taking it and using it to drive change to get you to a desired state or desired outcome.”
Too often, leaders try to change the system, when the focus should be on quickly changing the behavior of people, he said. If government wants more citizens to use a new passive data collection app or encourage more local startups to think about their communities, the predictive data engineer will use gamification to create incentives that are informed by data to make change happen fast.
“This role, I think, is about ingesting all this data but then using it in an automated, scalable, repeatable way,” Paharia said. “It can’t be teams of analysts looking at it and then figuring out what to do — that’s what computers are good at.”
Gamification becomes more powerful as more data points about individuals become available, Paharia said. If, for instance, it becomes clear that employees are driving to work during rush hour each day, this information can be used to create incentives for workers to drive at another time or to use a different route. If enough people are influenced, the impact is realized through reduced traffic, increased worker efficiency, reduced pollution, saved time and a happier community.
The dynamic of 21st-century roadways is changing. California, for instance, is preparing legislation for self-driving cars that could be in place as soon as 2015. Although it may take decades before autonomous vehicles hit the highways in large number alongside traditional vehicles, work is under way to make the technology roadworthy.
Edwin Olson is an assistant professor of computer science and engineering at the University of Michigan. His team placed fourth in a DARPA Urban Challenge, and his leadership of a winning team in a 2010 robotics competition led to the Department of Defense awarding him $750,000 for his work with autonomous robots. He’s now collaborating with the university and Ford on self-driving vehicle technology.
“We have a fleet of six 2014 Fusion hybrid cars that are driving autonomously on Ford proving ground areas and those will be driving in more places soon,” he said.
Just as governments today see little difference between trains and the rails they ride on, autonomous vehicles will one day be considered part of a city’s infrastructure.
How and when society will reach a world that is mostly self-driving cars will depend on society’s priorities, Olson said. Today’s fatality rate is one per 100 million miles driven, which is very good and will be “exceptionally difficult” to replicate with self-driving cars, said Olson. If society decides speed isn’t that important, then it becomes much easier to put self-driving cars on the road, he said. If convenience and quality of life gained by those who can’t drive becomes a higher priority, then a concession of more dangerous roads could also expedite adoption, he added.
One government role of the near future will be in fostering autonomous vehicle adoption, which could be accomplished by offering incentives like special magnetized highway lanes dedicated to self-driving cars, Olson said.
“It’s possible that cars might ultimately end up with a sort of flight-traffic control system that aircraft use, in which case there might be municipal managers doing things,” Olson said, although he considers it a long shot. Instead, he expects software will assume that role. “Even in the aviation space, that could be largely automated and if autonomous cars are accepted socially, then I would guess that automating a lot of that high-level management would also be socially acceptable.”
But even if software is put in charge of managing autonomous traffic flow, there would still be a human worker to adjust various thresholds for speed, safety and congestion levels, and to handle exceptions like rerouting traffic around construction. A highly automated system enabled by the Internet of Things would allow a city’s air- and land-based drones to co-exist with the city’s autonomous and manually driven vehicles, although human oversight would be necessary for when things go wrong, at least until true artificial intelligence is developed.
An autonomous vehicle manager would be responsible for identifying unknown objects in the city’s mapping interface and issuing citations to drivers who aren’t equipped with the appropriate autonomous vehicle stickers or to alert vehicle drivers who aren’t equipped with the proper sensors.
With work by Google and several large automakers advancing the self-driving vehicle space, many seem ready for self-driving cars to come soon, but Olson warns people to temper their expectations. The biggest challenge is getting an intelligent system to understand “weird scenarios,” he said. To a human, most weird scenarios aren’t weird at all, but easily explained through intuition. When a ball rolls into the street, a human driver will stop because he knows a child may soon run after it. They can hard-code those kinds of things into autonomous systems, Olson said, but the problem is that there are millions of one-off scenarios that machines continue to struggle with, but pose no challenge at all to the human brain.
A few years ago the Federal Aviation Administration (FAA) set a deadline of September 2015 for full civilian drone integration into national airspace. As civilian drones take the air, it’s natural that government agencies will follow.
Today, potential government drone use cases are popping up in new research projects weekly. There are drones that search for damaged pipes and power lines, drones that repair tunnels and potholes, and waterborne drones that check sewers for damage while monitoring chemicals in the water. Early experiments by several pizza companies and Amazon.com forecast that it’s only a matter of time before drones are routinely used for rapid air delivery. The FAA predicts there will be 30,000 drones in operation by 2020, many of which are sure to be government operated.
Maintenance of municipal drones likely will be outsourced to contractors, but managing the fleet will be government’s job. A drone fleet engineer will be responsible for monitoring what the drones are doing, directing them to new tasks, or sending people to physically retrieve the machines when there’s a problem. Real-time sensor data will show where the drones are, what they’re doing and what they see. As an integral part of the Internet of Things, drones will play an important role in a government’s predictive data capabilities.
As the world economy flattens, more agencies will turn to the public for manpower, products and ideas — especially as baby boomers leave the government workforce. The collective intelligence engineer will serve as the conduit who brokers the relationships between management and the outside world.
A recent informal NASCIO survey found that 40 percent of the public-sector IT workforce would be eligible to retire within the next year. In most states, more than 50 percent of public-sector IT workers are above the age of 50.
Not only are workers poised to leave, finding replacements isn’t easy. “The interest of those under 30 years of age to work in state government is very low,” said Doug Robinson, executive director of NASCIO.
As baby boomers head out the door, it is likely that government IT staffs will shrink. Technology as a service is supplanting the need for in-house workers, and increasingly intelligent software is replacing the need for menial human labor inside and outside of government. The human IT skills that are needed will be more often found externally, Robinson said, adding that the role of government will be not to provide talent itself, but to broker needed skills from the outside.
Collective intelligence engineers will be an integral part of government IT in five to 10 years, said Dustin Haisler, a former local government CIO who now is chief innovation officer for e.Republic (publisher of Public CIO). The Internet has transformed America from a consumer economy into a creation economy, and there’s a surplus of talent waiting to be tapped.
“Basically everything that’s happening in the sharing economy with Airbnb, TaskRabbit, Amazon Mechanical Turk, MobileWorks, all these organizations are all built around making human intelligence accessible in an HTTP request,” Haisler said. “It’s basically a protocol to access human labor.”
Even for traditionally internal government functions, agencies will turn to the public to find the skills and information they need. Projects like The Copenhagen Wheel, a device that turns bicycles into data collectors, foreshadow a future where everything people do will generate some form of usable data.
“You’re going to have this new role of someone who understands behavior and human intelligence and the whole aspect of connectivity that helps design systems to leverage people in ways to create value for communities and for cities,” said Haisler. “It could be as simple as environmental reporting or it could be as complex as, ‘How do we crowdsource a business process?’ The future workforce is going to be completely decentralized.”
The roots of a decentralized workforce can be seen in government projects today, particularly in work by the Boston Mayor’s Office of New Urban Mechanics (MONUM). An app like Street Bump, which uses citizen smartphones to passively monitor roads and identify potholes, is the kind of thing that will become more and more common in government, said Nigel Jacob, co-chair of the office.
The shift toward greater public cooperation means it’s important for government to build trust with its citizenry, Jacob said, and that can be difficult. One challenge for collective intelligence engineers will be appropriately reframing government problems so the public can address them. It’s an art, Jacob said, and all signs point to greater public involvement, not just with participation in government but also in direct decision-making.
A California-based startup called PlaceAVote is trying to reshape representative democracy by placing two candidates in political office who would act as a conduit to the public will, voting on issues based on the outcome of votes made through an app. In 2016, the company will attempt to place 20 more filler candidates in major cities around the country. The idea may fail, but the attempt alone shows that there’s a will to use technology as a connector between people and government.
MONUM Co-Chair Chris Osgood pointed to his office’s Public Space Invitational as another example of how government will partner more with the public to solve problems. The program asks residents for ideas to improve the city’s public spaces.
“The collective intelligence engineer will do for procurement what Wikipedia did for Encyclopaedia Britannica,” Osgood said. “It will totally disrupt the way in which government partners with entities to solve real, challenging problems that they face.”
As organizations turn to a decentralized workforce, and increasingly rely on machines that collect and feed on data, there’s another effort to make those systems intelligent enough to work without any outside help.
But before that happens, someone will have to hold the software’s hand as it learns how to replace everyone. A machine-learning engineer will drive productivity and efficiency to unprecedented heights, assisting artificially intelligent systems in completing the tasks they can’t yet do on their own.
It might sound silly, but many intelligent, successful people are taking the threat seriously. Tesla Motors founder Elon Musk told CNBC in June that he’s worried of a possible future that resembles the war-torn Earth seen in the Terminator movies: humans struggling daily to prevent their own extinction.
Many believe true machine intelligence will arrive in less than 20 years. Scientist and inventor Ray Kurzweil thinks the world will reach technological singularity in 2045. No one knows for sure when it actually will happen — if it happens at all — but machine intelligence is witnessing exponential growth today, said Martin Ford, entrepreneur and author of The Lights in the Tunnel: Automation, Accelerating Technology and the Economy of the Future.
The pool of jobs that robots can do is growing as they learn to do more sophisticated tasks. “Any job you can think of now where someone sits in front of a computer or a telephone and does the same kinds of things over and over again, that’s very likely to be automated 20 years from now,” Ford said. Automation is one thing, but most human jobs require at least some quotient of rationality, and that’s where machine learning will bridge the gap from replacing menial labor to replacing most kinds of labor.
A New York-based startup called WorkFusion created a platform that automates much of the work a middle manager might do. It can post work ads to job boards like Craigslist, analyze the skills of the respondents, rate them, and assign various tasks based on what’s most appropriate for their skill sets. Software today isn’t smart enough to do the kind of work the people being crowdsourced are doing, but that’s where WorkFusion’s other features set it apart.
“Initially, it automates the management level, and so far, that sort of sounds OK for jobs because it’s actually creating work for people to do under the direction of this AI system,” Ford said. “[But] as these people are working, it essentially learns from what they’re doing and it incrementally can automate their work further. As they’re doing things, they’re required to mark it up in ways that indicates to the machine exactly what they’re doing so that the machine can, over time, understand what they’re doing and further automate the whole process.”
A spokesperson at WorkFusion explained that the company’s software is the stepping stone for a new paradigm of outsourced labor. As the middle layers of the workforce disappear, organizational leaders will use platforms like WorkFusion, until software learns how to replace their jobs too.
“There are a lot of people who believe in a singularity,” Ford said. “There are a lot of people who think that by the end of the 2020s, we’re going to have machines that can think. I don’t know. To me, that’s highly speculative and it sounds a little bit aggressive, but anything is possible.”