Slowing the pace of technological change?

I share the concerns raised in Vogt’s commentary, ‘How Fast Should We Innovate?’ Controlling the pace of technological change is one of the epochal challenges of this era, and I offer suggestions to facilitate scholarly inquiry, collective deliberation, and public policy.

Two framing moves

Most writing on the subject of pace focuses on individuals’ and subcultures’ subjective experiences: John Dewey observed a ‘mania for speed’ (1927) long before Alvin Toffler discovered ‘future shock’ (1970) and nearly a century prior to Judy Wajcman’s STS perspective in Pressed for Time (2014). ‘We’re always chasing time,’ averred a sleep-deprived long-haul trucker, surveilled by bosses while at the mercy of nearly impossible schedules (Menzies 2005, 36). Energy-extraction boomtowns have long been recognized as socially dysfunctional (Freudenburg 1984) – and much of the world now resembles a boomtown. Contemporary commerce, communication, and transport are said to have generated a hyperculture, ‘a swirling vortex that today sucks into itself all elements of individual experience, thought and emotion’ (Bertman 1998, 84).

I am disposed to accept this general picture although I would prefer greater nuance in the claims – more acknowledgement, for example, that hours actually spent on work and housework have remained fairly steady (albeit unfairly distributed by gender and social class). And some of the technosocial disruption has been beneficial for some people – relaxing formerly overbearing constraints from marriage, religion, in-grouping, and social convention. However, my main quarrel with stories about the ‘no time’ problem is a classic level-of-analysis issue: preoccupation with micro-level symptoms distracts from study of the institutions and political-economic practices causing the difficulties. What is driving the pace of innovation, where are the potential brakes, and what would it take to selectively decelerate somewhat adroitly?

A second important reframing of pace-of-change thinking is to stop using the pronoun we, because we in fact rarely innovate – they do. Corporations with the highest R&D spending are based in the U.S. (11), Germany (2), Great Britain (2), Switzerland (2), France (1), Japan (1), and South Korea (1) (Strategy& 2016). California venture capital and Silicon Valley predominate among start-up firms globally; the U.S. military determines more than half of weaponry R&D; and those driving permissionless innovation (Dotson 2015) are disproportionately young, male, affluent, and white – with the blindered standpoints that come from a narrow demography.

Given that a majority of the world’s population has little to say over how fast the technology revolutionaries move, it is misleading to deploy careless collective nouns and pronouns that obscure lack of agency. We, humanity, and Americans hide, profound difference and inequality, fail to point toward latent or manifest social conflict, and implicitly paint a nonpartisan facade over an inherently contestable and partisan matter. If a relative handful currently drive the lion’s share of technosocial change, largely ignoring others except as potential customers, let us speak frankly about the fact.

‘Appropriate’ pace

For governance of pace to become conceivable, questions about ‘How fast is too fast?’ would have to become a lot more common. I presented a paper on excessive pace 30 years ago at The Society for Social Studies of Science, but this is my first published writing on the subject. I deeply regret the delay and take it as a personal failing; however, if a politically radical senior scholar shrinks from the topic, it probably indicates how hopeless many others must feel in the face of the juggernaut.

As intellectuals, fortunately, we need not calculate the odds of practical effect in order to decide whether and how to study a subject. To begin explaining what I find conceptually compelling about the pace of innovation, let me share observations from four mid-twentieth-century public intellectuals who worried about the problem. First, British science journalist Ritchie-Calder got me ruminating about the possibility of diminishing marginal social returns to innovation. Decades before cell phones and widespread Internet, he observed that ‘there is so much knowledge that we cannot cope with it effectively and so much technology that it tyrannizes us’ (Ritchie-Calder 1968, 112). Although a mere claim – and an overstatement at that – I found it an intriguing hypothesis.

A second provocation came from Ukrainian-American engineer and entrepreneur Boris Pregel, who eventually became president of The New York Academy of Sciences and observed that technological changes ‘affect all aspects of our daily life and the man in the street has great difficulty in assessing the consequences of this impact.’ Engineering commentator Edward Wenk went a step farther, proposing that ‘The swift pace of technological change no longer matches the response time of human affairs; technical prowess may exceed the pace of social skills, especially our ability to anticipate second-order consequences and take crisis-avoidance measures’ (1979, 4).

Finally, novelist and social critic Paul Goodman helped me distinguish between technoscientific study and its actual translation into usable artifacts and techniques. ‘The issue is not whether research and making working models should be encouraged or not,’ he said.

They should be, in every direction … .The point is to resist the temptation to apply every new device … . (New) ideas may be profitable for private companies or political parties, but for society they have proved to be an accelerating rat race. (Goodman 1969, 3)

Although these and other twentieth-century commentators pointed out that technological pacing poses deep conundrums, their writings did not much clarify either the mechanisms accelerating pace or the mechanisms potentially available for selectively slowing down. Fortunately for the scholar in me, political scientists were at about that same time beginning to examine decision making theoretically, moving away from conventional preoccupations with legislatures, executives, and political parties.

One small but crucial insight involved the human predicament of ‘small brain, big problems,’ which meant that no person, organization, or society is smart enough to make complex new decisions correctly via analysis. Study and calculation can help, of course, but stalemate in decision analysis can be avoided only by comparing a manageably small number of possible choices, pursuing one that seems more attractive or less noxious, and then modifying on the basis of experience (Lindblom 1959). Reliance on trial-and-error learning thus is inescapable, and figuring out how to improve learning in technoscience policy-making brought me to the problem of pace (see below).

Where millions of people are involved, choice and learning obviously become exceedingly difficult. But there is one important way that large numbers help: the potential intelligence of democracy rests on the fact that fewer considerations are likely to be overlooked as the diversity of influential participants increases. In principle, moreover, monitoring and error correction gain more eyes and ears and interpretive lenses – a collective application of standpoint epistemology.

Melding the above points, it became apparent that decisions about the pace and direction of potential innovations could be improved by enabling a greater diversity of stakeholders to engage in more effective trial-and-error learning, leading to iterative technological reconstruction. Among other requisites for that to be undertaken systematically were these:

• There must be time for problems to show up: it took some 20 years for asbestos workers to get lung cancer;

• Colleting and interpreting relevant data often requires funding, access, and expertise;

• Intelligible interpretations must reach stakeholders in a timely way;

• Proposing, deliberating, and negotiating changes typically requires venues where affected parties can interact with those in authority, together with low enough levels of technological momentum that changes are perceived as potentially feasible;

• Those with authority and/or agency must have incentives to engage in negotiation and technological reconstruction.

The implications for pace should be apparent. One is that the number of trials occurring at a given time will have a significant bearing on the capacity to monitor, diagnose, and correct errors: other things equal, more trials means less attention to each trial. For example, chemical manufacturers have overwhelmed the US Environmental Protection Agency’s chemicals testing unit by proposing several thousand new chemicals annually. Instead of several person years worth of attention, the tiny staff can devote only weeks or months of scrutiny to each premanufacture notification. Even more striking was a decision by overwhelmed EPA officials to simply ignore a provision of the 1976 Toxic Substances Control Act that required new study every time an approved chemical was applied to a new use. And most of the 80,000 chemicals in commerce have never been adequately tested.

Members of an intelligent civilization no doubt would increase funding for monitoring and learning. But sufficient haste can make effective monitoring impossible, as when 1000 mw nuclear reactors were designed and sold without waiting for operating experience with smaller ones. The industry likewise diffused in many countries before figuring out how to deal with nuclear wastes, and before ascertaining whether a high percentage of the public would find the technology acceptable. Suburbanization including malls undermined older businesses, contributed to white flight from inner cities, reduced the tax base available to fund urban services, and made the neighborhood café and pub impossible – it is apparent in retrospect (Dotson forthcoming). Had developers, planning and zoning boards, homeowners, and others proceeded more gradually, there would have been time to evaluate these ‘experiments’ while considerable flexibility remained (Collingridge 1980).

But rapid pace not only tends to obscure defects in original design and implementation, it also makes it difficult to perceive accurately whether recently initiated technologies are unfolding acceptably. Under the go–go regime, would-be opponents may not even understand what they should be worrying about; they have trouble marshalling evidence to win attention from media; and they are handicapped in envisioning alterations to propose. Even for highly salient technologies such as robotics and other automation threatening several hundred million jobs worldwide, it is difficult to galvanize debate, negotiation, and reconstruction before problematic innovations develop nearly unstoppable momentum and obduracy (Hughes 1994; Hommels 2008).

One category of technological pacing is especially worth noting. The overall ensemble of technologies restructuring twenty-first-century life has no clear constituency that might press for reconsideration. Only vague metrics are available by which to analyze any suspected problems, and there is no one exercising overarching authority to whom complaints or proposals could be directed. Nor is there a venue for deliberating proposed changes, and technological civilization’s incentive structures cannot readily motivate action on complexly interacting sets of pacing problems. In short, rapid pace is the enemy of intelligent trial-and-error learning both for specific technoscientific endeavors and for the overall pace of change.

Possible tactics

Some potentially effective levers for slowing pace actually are available, the equivalent of taking one’s foot off the accelerator. Without active governmental funding, many areas of technoscience-catalyzed innovation would slow down appreciably. If U.S. taxpayers presently are giving $150 billion to support R&D (AAAS 2016), perhaps they have a right to reduce the subsidy?

More than half the total is for the military, but DARPA and other military sources have been crucial for a good chunk of the major innovations over the past half century and more. Stemming in part from military funding are the Internet and many contemporary electronics, as well as AI and robotics; even environmental innovations such as photovoltaics benefited in important ways from military subsidy. Rapid escalation in the cost of medical care stems in no small part from research catalyzed by the National Institutes of Health, which now enjoys a budget of $33 billion annually.

Equally significant is a subtler channel through which taxpayers unwittingly stimulate innovation: The expenses of developing and marketing new products presently are tax deductible, including not only plant and equipment, but salaries for lobbyists, accountants, attorneys, salespeople, and factory workers. In addition, through investment tax credits, businesses can claim up to 40 cents on every dollar expended for certain kinds of innovation. Tax deductions and credits may be a good way of reducing the risks of R&D investments in high-priority fields where important needs depend on technical change. As an across-the-board strategy for pacing technology, however, tax credits and deductions are dubious devices because they have only forward gears and accelerator, no brakes.

Might it be possible to stimulate high-priority innovations while slowing down lesser efforts, thus making each trajectory easier to track and assess? In effect, the NSF and NIH processes for reviewing research proposals do this (by funding research ranked highly by peer reviewers). The system has been partially captured by the scientist-beneficiaries; nevertheless, a bewilderingly complex universe of forefront science is systematically winnowed and prioritized. Might that system offer lessons for prioritizing technological goals and means?

One need not envision a gargantuan bureaucracy determining what technologies are to be valorized. Recent experiments with ‘upvoting’ on Reddit, customer product ratings on Amazon, Kickstarting, going viral, and other Millennial-friendly sorting mechanisms provide hints of how millions more people could be drawn into decentralized conversations and decision making. I believe that what is now known as technology policy should come to be seen as unacceptably primitive – because it is so barely democratic, because it is disproportionately thumbs, and because it covers such a small fraction of the innovation technosphere.

No one now knows how to structure a sophisticated technology priority-setting process, and I would love to see STS-oriented theorists take on the challenge. But as a thought experiment to open minds to what may be a wide set of potentials, consider the possibility of allowing tax deductibility only for technological innovations that have been certified as high social priorities. Entrepreneurs would be free to pursue other innovations, but shareholders or venture capitalists would take all the risks, recouping their investment only if an innovation ultimately proved so successful that all development and other early costs are repaid via sales. This obviously would be a strong disincentive, possibly cutting the annual number of technological innovations by half or more. Attention thereby would be freed for improved monitoring, debate, and honing of the worthiest innovations.

The idea of course is outlandish by contemporary standards. Before jumping to endorse the status quo, however, one might pause to reflect on the fact that proposed restrictions on business usually have been greeted by fears, threats, imaginings. Thus the United Autoworkers Union opposed the 1978 ‘gas-guzzler tax’ on fuel-inefficient cars, arguing it would take away jobs manufacturing large cars; but U.S. manufacturers avoided the tax by adroitly altering engine sizes, vehicle weights, and product mix. A century earlier when Congress was debating the first pure food law, one irate legislator claimed that requiring canned foods to actually contain what the label said was just a step short of mandating table manners! Thousands of mandates and restrictions on techno-economic endeavors have been greeted initially with disapproval and outrage, then grudgingly accepted, and eventually taken for granted.

Some of these restrictions have had an effect on pace, of which the best known is the lengthy testing period for new pharmaceuticals. Complaints continue from would-be patients, physicians, economists, and conservative politicians, but the benefits of quick marketing are generally regarded as less important than a high bar for safety and efficacy.

Elected officials also accelerate and decelerate military RD &D programs in response to budgetary and foreign policy fluctuations. Design and manufacture of new weapons, transport, communications, and computer systems tend to move at an agonizingly slow pace from the perspective of military planners. The F-35 Joint Strike Fighter – the most expensive weapons program ever attempted at about $1.5 trillion – took recognizable form by 1993; yet Lockheed will only deliver the 2400 new aircraft in 2017–2026 and beyond. If all technological innovations took a quarter of a century, we would not be having this discussion; so the case illustrates that rapid pace is not inevitable, even for projects that some consider essential.

Conclusion

If no manufacturer, government regulator, or consumer would find acceptable an automobile without brakes, should a member of technological civilization endorse an innovation system that lacks institutions for slowing the pace of change when it becomes excessive?

I have suggested that the greatest contributions to social thought and collective action will not come from worrying about ‘finding time in a digital age’ or other efforts ‘to contest the euphorics of speed’ (Wajcman 2014, 184). For the psychocultural experience of rapid pace is symptom more than cause, and it is perceptual as much as tangible. Closer to the heart of the pace-of-technology problem is the incontrovertible fact that trial-and-error learning is much more difficult when change is rapid. Because trial and error is the main technique both individuals and organizations use to successfully reconstruct complex endeavors (when they do), I see no current substitute for directly opposing the institutionalized incentives and practices now driving rapid change. Better understanding of those forces is one task for scholars. Perhaps even more important would be social and policy analysis envisioning alternative political-economic and sociocultural institutions capable of reining in the runaway world. For activists cannot very well demand changes they cannot envision.

Recognizing that the issue of pace is inherently contestable and partisan, STS scholars and technoscience policy analysts cannot very well avoid asking themselves and each other: Whose side are you on? When thugs are victimizing a hapless family out for a stroll, the bad guys are going to win unless bystanders expand the scope of conflict by calling police or otherwise intervening. Roughly the same applies to technological pacing: It is clear in most domains who is winning, and assuming a stance of innocent bystander allows the dominant ones to continue dominating. Belief systems, funding, organization, and momentum are mostly on the side of the innovators. And with the exception of a few highly unusual arenas such as GMOs, expertise in the sciences, engineering, biomedicine, and technology policy tends implicitly or explicitly to support rapid pace. For STS scholars, a reasonable alternative to phony neutrality would be to use social analysis in a thoughtfully partisan manner to assist the underdogs.

In sum, although controlling the pace of technological change certainly is farfetched at this point, no one knows what may become possible. After all, hardly anyone foresaw the upheavals that ended Soviet communism, and few recognized in advance the difficulties the European Union would face regarding poor-versus-affluent member states and regarding open borders and terrorism. Whatever the unknowable odds against it, deciding upon and creating an appropriate pace of change is one of the central tasks that would have to be undertaken by a collectively wiser civilization. If new and modified institutions cannot be envisioned and empowered to selectively speed and slow the pace of innovation to match diverse needs and monitoring capacities in varying contexts, then a significantly improved civilization probably is impossible. Indeed, there is a good possibility that technological civilization will continue going further adrift in a ‘sea of unintended consequences’ (Winner 1977, 89).

Disclosure statement

No conflict of interest was reported by the author.

Notes on contributor

Edward J. Woodhouse is a Professor in the Department of Science and Technology Studies at Rensselaer Polytechnic Institute. He researches how political-economic institutions can improve the governance of technological civilization and analyzes science and technology policy, seeking wiser, fairer solutions that those now prevailing. He holds a PhD in Political Science from Yale University and a BA in Political Science from UC Santa Barbara.