Illustration by Amy Thompson
On March 12, 2014, Google called on the world to celebrate the 25th anniversary of the Internet, which was born, in the company’s view, when the first web browser was released to the public. Although the earliest Internet communication dated back to 1969, only those few with advanced technical skills could use it. With the arrival of graphical browsers, the Internet was opened to all of us, and Google, with help from early government investment, took off to become one of the richest corporations in the world. By 1993, the Internet was so widespread The New Yorker could publish a cartoon—it remains the magazine’s most viewed—featuring a dog sitting in front of a computer screen, explaining to a fellow canine next to him, “On the Internet, nobody knows you’re a dog.”
Even as the celebrations of the Internet’s adulthood were taking place, the so-called Next Internet was emerging from its infancy. Google acknowledged this when, in a revealing 2015 interview, Amit Singhal, a senior vice-president at the company, declared that its search engine—the innovation that catapulted Google into the global business elite—was now a “legacy” system, a euphemism for “still useful but destined for the trash heap.” The future for Google, other large firms and Next Internet startups alike, will be in the development of new forms of mobile-friendly search engines appropriate to the new ways we are using the Internet, said the executive.
It would be presumptuous to map out the precise composition of this next stage in the digital world, but one can conclude with some confidence that whatever the Next Internet becomes, it will likely dwarf its older sibling in the power to influence every facet of life, as well its potential to wreak havoc. So far, the Next Internet still bears some of the characteristics of the one born in 1989. But it is taking off rapidly and already challenging many of the characteristics associated with the original Internet, particularly its founders’ vision of a democratic, decentralized and pluralistic digital world.
Big data, the cloud, and the Internet of Things
The Next Internet brings together three major systems: cloud computing, big data analytics, and the Internet of Things. Combined, these systems promise companies and government agencies centralized data storage and services in vast digital factories that process and analyze massive streams of information gathered by networked sensors stored in every possible consumer, industrial and office device. The New Internet therefore establishes a very different set of standards that make the problems we have associated with the first Internet look comparatively minor. The environmental, privacy and labour challenges alone are monumental.
The original Internet was a set of networks that provided users with the software to locate information on millions of servers located all over the world and to connect with others through these server-based networks. It was decentralized to the extent that the computer servers anchoring the system were both large and small, public and private, and managed in a myriad of ways. The brilliance of the Internet was figuring out how to get this decentralized, distributed world of servers to talk to each other and users through a simple, universal software standard.
This structure began to change fundamentally with the growth of cloud computing, the first building block of the Next Internet. The cloud is a system for storing, processing and distributing data, applications and software using remote computers that provide on-demand IT services for a fee. Familiar examples include Google’s Gmail, the online storage company Dropbox, and Microsoft Office, which increasingly distributes its widely used word processing and business software online for a monthly fee.
The cloud enables businesses, government agencies and individuals to move their data from on-site IT departments and personal computers to large data centres located all over the world. This imperative to free up storage space opens a rapidly growing business for companies that profit from storage fees, from the services provided online, and from the sale of customer data to companies interested in marketing products and services or simply storing it until it becomes a valuable asset in a merger/acquisition deal.
What happens next to all of this valuable information remains literally and figuratively up in the air.
Government surveillance agencies like the U.S. National Security Agency (NSA) and Canadian Security Intelligence Service (CSIS) also work closely with cloud companies to meet their security and intelligence needs. The diverse collection of servers providing the foundation for the original Internet has evolved into a global system of data centres, each containing tens or hundreds of thousands of linked servers, operated primarily by private corporations and government military and surveillance authorities.
The leading science journal Nature made very clear the practical difference between the original Internet and one based in the cloud when it called on the U.S. government to establish a “Cloud Commons” for biological research, especially in genomics. It did so because research on large data sets is far easier and faster to carry out in the cloud than through servers based in university research facilities (a difference in project time alone of between six weeks for the cloud and six months for the old Internet).
Sparked by this research potential, and even more so by a massive advertising campaign to encourage individuals and organizations to move “to the cloud” (including high-priced ads during the 2011 Super Bowl), the idea of cloud computing is now familiar to most Internet users. You could say that if the New Yorker’s dog cartoon gave birth to the first Internet then the Next Internet began with the magazine’s 2012 ad featuring a sad-looking boy explaining to his teacher how “The cloud ate my homework.”
The cloud embodies the massive outsourcing of computing and related services like sales, accounting, customer relations, finance and legal matters to the companies that own and manage data centres. It marks a major step toward creating a centralized, globalized and fully commercial Internet.
The major cloud providers are almost all large corporations including familiar names like Amazon, the world’s largest cloud business, Microsoft, IBM and Google. Through service contracts, most of these are well integrated into the military, intelligence, and surveillance arms of government. Amazon, for example, provides cloud computing storage and services for both the Central Intelligence Agency (CIA) and the NSA. Meanwhile, government agencies demanding heightened levels of security are building their own cloud facilities: in 2015, the NSA opened one of the world’s largest in a secure mountain location in Utah.
Cloud centres are global factories working on the raw materials of data to yield digital products (e.g., user profiles) and services that generate new data. In addition to the value of rationalizing and outsourcing corporate and government operations, supporters of the cloud point to the growing power of big data analytics to make the cloud even more attractive. This second leg of the Next Internet is made up of the tools used to analyze, package and sell data stored in the cloud.
In spite of the proliferation of fancy new titles that fuel enthusiasm for big data—data science professional is a favourite—there is very little that a social scientist would find especially novel. The process involves taking a large (often massive), almost always quantitative data set and examining the specific ways the data do or do not cohere or correlate in order to draw conclusions about current behavior and attitudes, and to make predictions.
Facebook, for example, takes the data generated by its 1.3 billion users and relates the “likes” associated with posts about celebrities, companies and politicians to views about society or commercial products (and, of course, cats). These enable the company to develop profiles on its subscribers that are then sold to marketers who are able to target Facebook users with customized ads sent to their Facebook pages. Google does the same for search topics as well as for the content of Gmail, and Amazon creates profiles of its users based on searches and purchases on its site.
Given the limitations of quantitative correlational analysis, especially the absence of history, theory, and subjectivity (qualitative data is ignored or poorly translated into numbers), we should hardly be surprised by big data’s failures; for example, on such projects as seasonal flu predictions and building models for economic development. These failures are mounting as quickly as the opportunities to manipulate online data for profit. Nevertheless, for simpler problems such as determining the likes and dislikes of every conceivable demographic cohort, the massively large stores of data available for analysis in the digital factories that make up the cloud offer major incentives for companies and governments to invest in both cloud data centres and big data analysis.
The Internet of everything
The cloud and big data are enhanced substantially by the growth of what is called the Internet of Things (IOT). From watches that monitor blood pressure to refrigerators that tell you to order more milk, from assembly lines “manned” by robots to drones that deliver what robots produce, the IOT promises a profound impact on individuals and society.
The IOT refers to a system that installs scanning and processing devices in everyday objects and production tools, connecting them in networks that gather and use data on their performance. We refer to the admittedly awkward term Internet of Things because, unlike the Internet we know, which connects people, the IOT primarily links objects. The sensors in a refrigerator form a network of things that report on what’s inside and how it is used. The IOT is made possible by advances in our ability to miniaturize scanning devices and provide them with sufficient processing power to permit the ability to monitor activity, analyze usage and deliver the results over electronic networks.
A 2015 report from the private think-tank McKinsey concluded that, by 2025, the IOT will have an economic impact of between US$3.9 and US$11.1 trillion ($5.1 and $14.6 trillion), which, at the high end, is over 10% of the world economy. Even discounting for the hyperbole that often accompanies tech forecasts by research organizations that are looking to drum up business from the industries they examine, the report is interesting in highlighting the likely impacts and identifying the organizations most likely to be affected.
Interestingly, according to McKinsey, the manufacturing sector leads the way for IOT as robotics and overall surveillance of operations enables more tightly managed and efficient factories and global supply chains. But these will also extend to offices, retail operations, the management of cities, and overall transportation as automated vehicles take to the streets and highways. Heightened monitoring will also take place in the home, with promises of greater control over heating and cooling, ordering food and supplies, and to the body as well, where sensors will continuously monitor fitness and body functions like blood pressure, heart rate and the performance of vital organs.
This sounds futuristic. And depending on your point of view either dystopian or utopian. But it speaks to the potential power of the new technology and to the fundamental differences between the original Internet and its successor.
Companies have been quick to take advantage of their leading positions in the digital world to rush into the IOT. Prime examples include Google’s research on the driverless car, the Apple Watch, and Amazon’s embrace of robotics in its warehouses to speed the work of order fulfillment, the use of drones for delivery, and new forms of packaging, including installing pushbuttons on products that automate the ordering of refills. The IOT has also given new life to an old industrial firm, General Electric, which was remade in the 1990s by shifting from manufacturing to finance. GE has now all but abandoned the increasingly regulated world of banking only to emerge as a dominant player producing devices essential to the IOT and making use of them in its own industrial processes.
Along with the benefits to corporations, the IOT holds out great promise for the military. In fact one reason why the Pentagon’s Defense Advanced Research Projects Agency (DARPA) is so keen to develop IOT is because it greatly strengthens opportunities to automate warfare through robotics and drone weaponry in addition to enhancing overall control over troops.
One enormously valuable result of monitoring every device and connecting them in a global grid of objects is the exponential growth in commercially useful data. Today, according to a Cisco report, only 1% of the world’s objects are linked, so the big promise of the IOT remains just that—a promise, yet to be filled. Nevertheless, it is forecast that by 2020, 50 billion connected devices will join the Next Internet, gathering and reporting data all the time. Massive new data sets will require both massive new cloud data centres and widespread use of data analysis to succeed.
“Currently, most IOT data are not used,” says McKinsey. “For example, on an oil rig that has 30,000 sensors, only 1% of the data are examined. That’s because this information is used mostly to detect and control anomalies—not for optimization and prediction, which provide the greatest value.”
How to use this data, internally and as a marketable commodity, is one of the biggest challenges facing the IOT industry.
The corporate-military-data complex
Most of what is written about the Next Internet is technical or promotional, emphasizing the engineering required to build it or touting its potential in sometimes accurate, sometimes dreamily hyperbolic terms, from creating a world of nonstop leisure to ending capitalism as we know it. We are just beginning to see some discussion of the serious policy issues that arise in a world of massive data centres, constant analysis of human behavior and the performance of objects, and ubiquitous connectivity. These include the concentration of power over the Next Internet in a handful of mainly U.S. companies and the military-intelligence apparatus, the environmental consequences of building and maintaining massive data centres and powering systems, threats to privacy and security, and the impact of automated systems on human labour.
A handful of familiar American companies, currently led by Amazon, dominate the Next Internet. The company was among the first to build a one-size-fits-all cloud service that attracted individuals and organizations (including Netflix) with its simplicity and discount prices, which dropped even further when Google and Microsoft began to challenge the online retailer’s frontrunner status. Indeed some have claimed that Amazon and it competitors have engaged in the not-so-fine art of predatory pricing by charging below cost for cloud services and compensating with above-market prices in other businesses where they enjoy market power.
Facebook and Apple round out the list of firms that used their power in social media, software and hardware in the original Internet to become leaders in the Next Internet. Legacy companies like IBM, Oracle, HP and Cisco have scrambled to replace their expertise in servicing (now disappearing) IT departments and pivot to the new digital world. However, the need to cannibalize old systems and remake their organizations has made the going slow.
In addition, there are firms that specialize in one or another of the constituent Next Internet systems, such as Rackspace and Salesforce.com, but these are constantly undermined by encroachment from the dominant companies. A wildcard in the Next Internet arena is General Electric, which is betting heavily on reinventing factories with IOT systems.
Two things stand out in the early configuration of the Next Internet industry. First, it is already highly concentrated and, as we’ve seen, dominated by American firms. Historians of technology will recognize the similarity of this pattern to the early days of electrification, telegraphy, telephony and broadcasting. In each of these industries, regulation and outright state ownership were required to control abuses and increase access at affordable rates. However, these remedies are less likely to be applied in a world where regulation and government ownership are no longer in favour.
A second feature, also mentioned briefly, is how dominant firms are benefitting from their close ties to the military and intelligence communities, providing them with Next Internet services and co-operating more often than not with government requests for information on users. In fact, close ties to the Pentagon, including its well-funded research arm DARPA, the NSA, and the CIA helps to explain why there are no challengers to U.S. hegemony over the Next Internet coming from Europe, whose telecommunications companies once led the world.
As for Canada, the government’s decision to build an oil-based economy and diminish support for technology and science-based investment means that the country has no successors to the once powerful Nortel or the pioneering Blackberry. Canadian companies do little more than fill small niches in the new systems that make up the Next Internet.
The only serious challenge to U.S. domination comes from China where the government has invested heavily in Next Internet technologies and made them a part of its overall policy and planning. This has benefited leading companies like Alibaba, Baidu, Huawei and Tencent, among others. Signalling its intent to challenge America’s lead, Alibaba has set up shop in Silicon Valley and, like other Chinese firms, is building on the enormous domestic market to extend its reach internationally.
While the prospects for China’s success are uncertain, it is interesting to observe that it is the only formidable force in the world standing between Silicon Valley’s giants and complete domination over the Next Internet. A look at the remaining policy issues reveals why this is so important and why it is essential that societies begin to consider the need for public intervention.
The still material world
Because the digital world is made up of invisible electrons zipping through the air, there is a tendency to see it as immaterial. Nothing could be further from the truth and the sooner this is recognized the more likely the environmental problems associated with the Next Internet will be addressed. Some of these problems are specifically related to the cloud and others are of more general concern. Cloud data centres are hardly immaterial structures and, as they come to fill the world, there are numerous emerging environmental policy issues.
First, there are question of siting and building data centres. Cloud companies typically locate outside urban areas and demand relief from property taxes. Because they are energy hogs, they also look for discounts on electricity. Their lobbying and bargaining power, especially with the local governments that often handle such issues, have contributed to their success. Breaks on energy charges are especially significant because the thousands of servers in cloud centres make enormous demands on electricity grids for power and cooling.
Moreover, cloud data centres often require relief from environmental controls because customer demand for nonstop service requires several layers of backup power that are often polluting (primary backup uses diesel generators) and sometimes carcinogenic. Many data centres require large, continuous supplies of water for their cooling systems and this requirement alone raises serious policy issues in places like the western United States where years of drought have taken their toll.
Some companies have responded to opposition from environmental groups, especially Greenpeace, by incorporating solar and other sustainable forms of power into their data centre power supplies. But as power requirements grow beyond the 3% of the global supply they now consume, systematic regulation is required, including a broad review of discount power deals.
Notwithstanding any progress in this area, the primary source of power consumption in the Next Internet lies elsewhere—in the sensors embedded in what is expected to be billions of connected devices, and the communication systems that connect people and things through cellular and other wireless networks.
A world of ubiquitous, always-on connected devices has energy companies salivating, especially the lobbying arm of the coal industry, which sees the Next Internet as an opportunity to build on what a study for the U.S. National Academy of Sciences calls “the renaissance of coal.” As one industry-sponsored report concluded, “The inherent nature of the mobile Internet, a key feature of the emergent cloud architecture, requires far more energy than do wired networks…. Trends now promise faster, not slower, growth in ICT energy use.”
By the standards anticipated in a digital world where the IOT is fully developed, today’s Internet is far from creating a connected world, let alone the singularity that fills the dreams of Internet enthusiasts. Only about 40% of the world’s population now has access and, as one might expect, it is concentrated in the developed world and in urban centres. And as mentioned already, we are even further away from the promise of ubiquitous links among objects with only 1% connectivity.
But even at this relatively low level, technical problems and criminal hacking plague the system. On one day in 2015, the entire U.S. fleet of United Airlines was grounded, the New York Stock Exchange shut down for several hours, and the Wall Street Journal’s computers stopped operating. All of these were explained as a result of technical “glitches.”
Just as this calamity hit the news, the U.S. government reported that hackers had stolen the personnel records of 22.1 million people including federal employees, contractors, and their families and friends who provided information for background checks. The haul also included 5.6 million sets of fingerprints.
Even with the current limited state of connectivity, similar incidents are a regular feature of the emerging Next Internet and even supporters worry that security issues may slow its development. Privacy and security concerns rise exponentially as greater connectivity increases opportunities for technical breakdowns and criminal hacking. One tech journalist referred to the IOT as “the greatest mass surveillance infrastructure ever.”
The all-seeing CEO and surveillance state
Notwithstanding new hacking opportunities, the most significant threats from the Next Internet arise from data-hungry businesses and governments. After all, the greatest attraction of ubiquitous computing is in the data collected on the behavior of people and the performance of objects. These offer opportunities for an enormous expansion in both surveillance capitalism and the surveillance state, with businesses refining targeted advertising and product development well beyond the crude systems that even today’s Internet makes possible, and governments deepening tracking and control of citizen behavior and attitudes.
Consider the commercial benefits to insurance companies that will be able to continuously monitor the health of customers, their driving habits and the state of their homes; or to governments that can adjust benefits and other services based on citizen behavior registered in their actions, as well as their interactions with one another, and with the things that fill their lives.
As GE’s sensor-equipped light bulb demonstrates, individual CCTV cameras will fade in significance as any basic device scanning public and private places has the potential to become a surveillance tool. Discussions of anticipatory selling, as well as anticipatory policing (euphemistically called “predictive analytics”) are worrisome to privacy advocates because they are attracting great interest from businesses and governments.
The impact of the Next Internet on jobs and the nature of labour is also an important policy issue. At first glance, it is tempting to think “here we go again,” since the impact of technology has been discussed for many years, but especially since the end of the Second World War, when the computer scientist Norbert Wiener generated considerable public debate by raising the spectre of massive job loss due to automation.
The Next Internet is creating and will likely continue to create jobs, including traditional construction jobs in the building of global networks of data centres. The new profession of data science—those trained in the analysis of large data sets—will grow. And the industry will need people skilled in the control, maintenance, and monitoring of networked things, though it is important to approach the impact of computer technology on jobs and the economy with caution. As Doug Henwood’s research documents, overall employment is much more closely connected to GDP than it is to computerization and, except for the late 1990s, when there was massive investment in hardware, the long-promised productivity gains from investment in IT have failed to materialize.
However, it appears very likely that this time around there are far more opportunities for the Next Internet to eliminate human labour, including professional knowledge work. In fact, one expert consultant prefers to define cloud computing as “nothing more than the next step in outsourcing your IT operations.” This is in keeping with a general tendency, which one researcher for the major consulting firm Gartner Associates summarizes succinctly: “The long run value proposition of IT is not to support the human workforce—it is to replace it.”
The Next Internet creates immediate opportunities for companies to rationalize their information technology operations. Again, from Gartner, “CIOs believe that their data centres, servers, desktop and business applications are grossly inefficient and must be rationalized over the next ten years. We believe that the people associated with these inefficient assets will also be rationalized in significant numbers along the way.”
Next Internet companies maintain that their systems can break a pattern in business organizations that began when the first large computers entered the workplace. Almost all business and government agencies insisted that it was essential to operate their own IT departments and, for the larger organizations, their own data centres. Next Internet supporters insist a few large data centres can meet this demand at lower cost with far fewer professional personnel. This process has already begun and early studies demonstrate that, even with limited downsizing of IT departments, companies are saving between 15% and 20% of their IT budgets.
The Next Internet also makes possible the widespread rationalization of all knowledge and creative labour because the work of these occupations increasingly involves the production, processing and distribution of information. According to one observer, “In the next 40 years, analytics systems will replace much of what the knowledge worker does today.” A 2013 report concluded that 47% of the current U.S. workforce is directly threatened and in the high-risk category for job loss.
Whatever the precise share under threat, there is no doubt that the current trend is to move an increasing portion of the work that knowledge workers perform to machine systems, specifically through software that can carry out tasks that once required human intelligence. Analytical systems at the core of the Next Internet are taking prominence in education, health care, the law, accounting, finance, sales and the media. Thanks to the Next Internet, both private and public sector organizations are encouraged to outsource all but their core business processes to companies like Salesforce.com, which specializes in managing vast databases of customer information, a function that traditional marketing and client service departments within organizations typically performed.
Flexible labour at the breaking point
The expansion of outsourcing to computers raises serious questions for the entire global system of flexible production. According to Gartner, “That outcome will hit all economies—especially emerging ones like India that now dominate technology outsourcing.” It may be an overstatement to declare, as did Forbes magazine, “We are all outsourcers now,” but it certainly makes feasible more kinds of outsourcing: “Outsourcing is no longer simply defined by multi-million-dollar mega-deals in which IT department operations are turned over to a third party. Rather, bits and pieces of a lot of smaller things are gradually being turned over to outside entities.”
Amazon is a leading force in this process with its Mechanical Turk business that charges individuals and organizations who wish to outsource micro-tasks to a worldwide reserve army of online piece workers. Combined with the promise of product warehouses full of robots to locate, pack and ship goods, and drones to deliver them, Amazon is the leading edge of the Next Internet’s push to expand labour commodification throughout the world. Reports on the Next Internet are full of talk about the opportunities to radically reorganize and manage global supply chains.Top of Form
What can be done to address these problems? First and foremost, it is essential to view them as intrinsically social and not merely technological in nature. The dream of ending capitalism with technology remains a fantasy. While technology plays a role in addressing serious policy issues, there is no simple digital fix to solve them. It will take concerted political action to break up the concentrated corporate power that is now making the Next Internet primarily a tool for profit.
It will also take global social movements, stronger versions than those that called for a New World Information and Communication Order in the 1970s and ‘80s, to build a digital commons for the 21st century. Furthermore, there is great need for social policies that make protecting the environment and sustainability central to all decision-making about the Next Internet.
We also require social policies that recognize privacy as the human right of access to the psychological space essential to developing individual autonomy. This means that privacy is an essential right of citizenship and not a tradable commodity. Protection of that space from commercial and government surveillance must be central to the choices made about the Next Internet.
Finally, we need social policies about employment and income that address the state of human labour in an age when automation threatens jobs, including now those of the white-collar workforce. Does this mean we should reopen the discussion of a guaranteed annual income? What is the right balance of such a policy with job creation? How can we facilitate organizing digital workers who tend to be employed in the “gig” economy of precarious jobs? Is the union at Gawker—a pioneering web-publishing success story—a good model for the future?
The digital world is at a critical juncture represented by two clashing visions. The first imagines a democratic world where information is fully accessible to all citizens as an essential service. In this world information is managed through forms of regulation and control that are governed by representative institutions whose goal is the fullest possible access and control for the greatest number of citizens. Governance might take multiple forms, including different combinations of centralized and decentralized approaches at local, regional, national and international levels.
The second vision imagines a world controlled by global corporations and the surveillance and intelligence arms of national governments. Under this model, the market is the leading force shaping decisions about the production, distribution and exchange of information, and corporations with market power hold the most influence. In this fundamentally undemocratic world, digital behemoths share power with governments who make full use of digital technologies for surveillance, control and coercion.
Fifty years ago, long before the first Internet, the Canadian scholar and policy analyst Douglas Parkhill chose the democratic vision in his book about the need to create a global system of computer utilities that guaranteed public control and universal access. Social movements had helped to tame private monopoly control over earlier essential resources like water and electricity by turning them into public utilities. Parkhill made the case that information was no less essential and no less in need of public control. The Next Internet presents us an opportunity to build on his vision.
Dr. Vincent Mosco (PhD Harvard) is Professor Emeritus of Sociology at Queen's University where he was Canada Research Chair in Communication and Society and head of the department of sociology. His most recent book, To the Cloud: Big Data in a Turbulent World, was named a 2014 Outstanding Academic Title by Choice: Current Reviews for Academic Libraries. Dr. Mosco is working on another book about the Next Internet, big data and the cloud, which will expand on ideas presented in this article.