“Disruption” gained credence as the term of art for transformative change in the Internet economy, with “incumbents” providing rhetorical contrast. Thus, the battle was joined between blinkered incumbents and bright-eyed disruptors. Billions of dollars flowed into Silicon Valley. Books were written. New business leaders were placed on pedestals. By September 2016, for the first time in history, technology companies were for many weeks the five largest companies in the world (Amazon, Apple, Microsoft, Google, and Facebook) measured in equity value. The disruptors had joined the incumbents.
There was just one thing these former disruptors forgot — for every action, there is an equal and opposite reaction. And that reaction can be less predictable than disruption itself.
Reactions to some pernicious effects of Silicon Valley’s “disruption/incumbent” dialectic have arrived in full force as 2017 dawns. As a recent essay by Brad Stone in BusinessWeek outlines, the rise of Trumpism and nationalism generally has come about partially because of a number of flaws in the Internet economy as currently realized:
- the economic inequality created by disruptors, who are more educated and urban, and the resulting resentments and insecurities motivating populist movements among the less educated and rural
- a “neutral platform” approach, with no human intervention or barriers to entry, which allowed Russia and some fringe cultural groups to foment resentment in the US, UK, and elsewhere by mixing warped realities with emotional appeals to create a fog of war around facts
- an information economy that counts every click equally, which allowed trolls and miscreants to undermine cultural norms, political contests, and stable sources of information while selling ads and making money
For good reason, the effects of the 2016 information economy provoked concerns about how Internet information companies have been conducting themselves and what they’ve been focused on, concerns that reached an inflection point with the shocking outcome of the US presidential election. As Stone writes:
Trump’s election was the ultimate wake-up call for Silicon Valley. In a way it largely wasn’t before, the tech community in 2016 was asked repeatedly to grapple with its broader role in society — not only with the secondary effects of its products but also its own easy characterization as a villain in an age of displacement and stagnation. . . . This was most evident in the dust-up over fake news, which revealed once again the flaws in the thinking behind “neutral platforms,” which treats all content equally as long as users find it engaging.
The tech industry and its adherents were caught off-guard, as they have been trapped in the “disruptor/incumbent” dialectic, frozen in rhetorical amber as others evolved to exploit their vaunted and profitable “neutral platforms.”
The economic benefits of “neutral platforms” came down to “all of the benefits, none of the responsibility.” The political risk of this unrealistic bargain was a ticking time bomb.
The “neutral platforms” notion informing modern information economics unraveled last year for scholarly publishers, as well. Many publishers happily noted record-breaking monthly traffic at least once this past year, only to realize too late that it was Sci-Hub scraping all their PDFs away. There are now such things as too much traffic and the wrong kind of traffic. Refining these measures will be a chore for many people and companies over the next months and years. The illusion of “neutral platforms” has been broken, as has the notion that all traffic is good traffic.
To make “neutral platforms” work, Silicon Valley came to rely upon algorithms and other technological solutions that, they believed, eliminated the need for human editors, reviewers, or curators. Job cuts became part of the new information economy’s growth, including Facebook’s notorious and ill-timed elimination of its human editors late last summer. The bet was that editorial or human decision-making could be effectively replicated in algorithms, while engendering less controversy. It didn’t work well. Fake news erupted on the platforms. Facebook is now knee-deep in controversy and backlash.
There were other warning signs that tech companies were playing with fire in their blind adherence to algorithms and their reflexive dismissal of human interventions:
- Uber charging “surge pricing” during a terrorist attack in Sydney
- Google’s search algorithms allowing Holocaust denial sites to dominate results for “did the holocaust . . .”
- Kellogg’s ads running atop racist and sexist Breitbart stories without the company’s knowledge
- YouTube running Miller beer ads over ISIS recruitment videos without the company’s knowledge
How did these things happen? Each algorithm saw data it was designed to respond to, and proceeded without any human intervention in place to prevent a short-circuit. In every case, the outcome was awful. In every case, humans saw what was wrong immediately, but had no direct way of fixing the problems, and responses took days or weeks, not minutes or seconds. The algorithms were in control.
It goes deeper. If Silicon Valley hadn’t been resting on its laurels, its might have seen that some of its first principles of disruption were becoming irrelevant. As a recent column in the Economist points out, the management principles informing disruption itself may now be badly outdated and incorrect:
- Principle #1: Business is more competitive than ever. This no longer seems to be true. In fact, business seems to be in a long arc of consolidation and rearranging of deck chairs, especially since the meltdowns of 2007-8. Since 2008, mergers and acquisitions have averaged 30,000 deals per year. Consolidated companies have become more profitable. As noted above, Silicon Valley itself is now dominated by mega-companies that are mainly involved in acquiring startups. If competition is about who can become bigger and less prone to disruption, there may be a case made. However, if competition is about the little guy overturning or disrupting the big guy, think again. In our industry, consolidation is just as rampant. We all remember the shock we felt when Mendeley was acquired by Elsevier. Now, consolidation in our industry is the norm — from Wiley acquiring Atypon to HighWire acquiring Semantico, now we’re seeing consolidation move from publishers to platforms. It shows no signs of ending there.
- Principle #2: We live in the age of entrepreneurialism. For disruption to occur, you need entrepreneurs. The Internet Age seemed to promise an explosion of new companies, and there are notable new entities, to be sure. However, the rate of business creation has declined since the 1970s. Recently, more companies have died than have been born. Tax laws punish new companies that have more than, say, 50 employees, stifling growth and feeding consolidation. And entrepreneurs are often looking for the “exit,” which mostly involves acquisition by a larger company. This doesn’t feed disruption; it feeds expansion and consolidation.
- Principle #3: Business is getting faster. Disruptors are thought to be more agile, quicker on their feet, and this increased speed and facility allows them to outwit the slow-footed incumbents. But consolidation is making for fewer customers for new ventures, and the larger companies are often caught in bureaucratic red tape. New layers of bureaucracy (audit, legal, compliance, privacy) lead to dithering that slows disruptors, as desires to manage cash flows and cash reserves during times of uncertainty dominate decision-making among their potential customers. Disruptors are caught in the resulting molasses. Stashing cash has taken over as a business priority, not innovation or disruption. Think of how little the pecking order has changed in the last decade. You could argue that we are stuck in 2007.
- Principle #4: Globalization is inevitable and irreversible. It became a management tenet that demographics and trends toward globalization would swamp efforts to blunt change. However, recent politics has shown that globalization is not a safe assumption and that demographics are not destiny. Both can be thwarted or reversed. From Brexit to Trump to TPP, globalization is under assault, and this assault is looking to be increasingly effective. Politics has not had its last say in how businesses grow and adapt.
Instead of disrupting the big players, modern information economics have disrupted the mid-tier, leading to haves and have-nots. One example of Silicon Valley economics in our market is open access (OA), which was born inside the first Internet bubble economy. As in other parts of the information economy (books, television, movies, newspapers), OA has disrupted the middle of our market, while feeding the trends outlined above — consolidation, conservatism, and stagnation.
The Internet once represented exciting potential for online publication when the economics were more demand-side. Now, it has become tainted by a grim grind of article processing and journal proliferation in the service of supply-side economics in publishing — more focus on authors than on readers, and more focus on quantity over quality. The recent battle over APC prices and efforts by Elsevier and Springer to prevent their disclosure show how unimaginative, consolidated, and conservative OA publishing has become.
In our industry, the disruption has occurred in ways nobody expected, behaving more like “collateral damage” than disruption. Examples of collateral damage include the struggles of many non-profit society publishers and university presses. There was a time when it was thought that the Internet might level the playing field for these entities, and allow their higher-impact and more relevant journals to thrive. Instead, these societies and publishers are more reliant than ever on contracts with large multi-national publishers, as the Internet Age proved to be too disruptive — expensive, complicated, and consolidated — for many of them. Rather than benefiting from increased competition, they’ve suffered from decreased competitive leverage. They survive in name, but under different conditions, and precisely the conditions the disruption hoped to avoid — under contract with the largest entities in the space, who are only getting larger.
For Silicon Valley and those inspired by its existing ethos, the future is clouded. A recent rant by Kara Swisher in Recode entitled, “Hell is Silicon Valley people who won’t grow up,” captures a number of themes more realistic innovators need to think about, as innovation is no longer as much about technology or disruption but about positioning and purpose. As she quotes Pejman Nozad in her piece:
Silicon Valley is still a place of big minds chasing small ideas.
From photo-sharing to link-sharing services, Silicon Valley has rearranged many value chains, but without much responsibility for the unintended consequences. Journalism weakened? Not their problem. Small towns thrown into poverty? So sad. Terrorists given worldwide megaphones for radicalization? The tech teams just run a service. Automation devastating worker pay and eliminating jobs? Get over it. A widening divide between rich and poor? Their Capitol Hill lobbyists are more concerned with future earnings.
An endless pursuit of technology without a concern for the potential downsides, and actual insulation from these downsides, is, as Swisher writes, starting to boomerang on its purveyors in the form of broad political backlash. As she writes:
It’s often referred to as a Peter Pan mentality, in which its denizens are trying to remain forever young in a land of perpetual boyhood, making things like photo apps and social media and new ways to play old video games. Personally, I think there is a far more sinister comparison to another fairy tale, that of Pinocchio’s transformation into a jackass on Pleasure Island. It’s a place where boys are indulged with endless fun until it becomes clear that there is actually a price for all that indulgence.
One price we pay is our grasp on reality. Even fact-checking sites like Snopes can’t battle against the tide of confirmation biases and misinformation swamping us, as a recent New York Times story outlined. The compulsion to view the world in a pre-established way is so strong among some zealots that editors at Snopes are not only being smeared on the Internet for fact-checking stories, they now fear for their lives:
Smearing people just because you don’t like what they’re saying often works to shut them up. But at Snopes you learn to grow a thick skin. I will always push back. At least until someone shows up at my workplace and kills me.
A related price we’re paying for treating disruption as desirable — via “neutral platforms,” information abundance, and unfiltered access — could be a more superficial and uncertain scientific and scholarly literature. As a recent post on Forbes outlines, the overall quality of scientific discourse could be dropping because quantity, velocity, and productivity are reinforced above all else in the modern information ecosystem. What seems to have been disrupted is intellectual rigor, the kind of thinking that occurs when you have to painstakingly and slowly work through a set of intellectual outputs.
This is not a minor point. As people become conditioned to expect information to just rise up to greet them instead of requiring the work and effort to suss out truth from fiction, we see an intellectually lazy result — fictions supplanting truths, yes, but worse, superficial or transitory information clouding deeper truths and more subtle trends.
Reading a full paper rather than an abstract will typically raise concerns, questions, and ideas a few moments with the summary won’t reveal. Deep reading is about interrogating information more intensely, while spending time contemplating the issue long enough to tease out new aspects, bring related knowledge or thoughts to mind, or develop expanded lines of inquiry.
Rapid-fire information accession deprives us of these deeper processes and keeps our superficial mental processes dominant — confirmation bias, anchoring, and availability errors, among others. A quote attributed to Dov Seidman, CEO of LRN, used on Kara Swisher’s recent Recode Decode podcast, sets it out well:
When we hit “pause” on technology, it stops. When we hit “pause” as humans, we start.
Whipped into a constant information frenzy, our processing units are constantly fighting to keep up, and more often than not, they are being defeated. As Maggie Oarth wrote in the Washington Post:
In a dramatic historical shift, too much information on the Internet — rather than a lack of information — has become the prevailing form of censorship. People cannot easily assess facts while doubts, lies and bias-confirming opinions are constantly whispered in their ears.
The possibility that academic research is being built on sand seems stronger with each passing year. The Forbes essay by Kalev Leetaru talks about the insidious habits Google Scholar has fomented among busy academics, who are working faster and faster to write papers and get published in order to maintain their career trajectories. In one passage, he discusses being confronted with the results directly:
Not a day goes by that an academic paper doesn’t pass through my inbox that contains at least one claim that the authors attribute to a source it did not come from. I constantly see my own academic papers cited as a source of wildly inaccurate numbers about social or mainstream media where the number cited does not even appear anywhere in my paper. Indeed, as but one example, I recently attended an academic conference where my Twitter Heartbeat paper was cited by one presentation as the source of a claim that 90% of all tweets include precise GPS coordinates. The percentage 90% does not even appear anywhere in my paper in any context and the actual conclusion of my paper was that around 2% of all tweets have precise geolocation information. When I asked the authors afterwards about this, they initially claimed I was mistaken and only after identifying myself as the author of the paper in question did they suggest that perhaps they had misread my paper or gotten the figure from a different paper and copy-pasted the wrong citation. Yet, this was not a one-off – I see this kind of error every single day.
The more insidious aspect of this is how academic publishing and academic life in general is being turned into more and more of a sport, complete with the circular logic of these endeavors. A study making the rounds claims to demonstrate that people who do well at academic publishing early have better careers later. That smacks of the circular logic of sports — early success translates to later success — which has been debunked but persists in the world of confirmation bias. The title of the InsideHigherEd coverage underscores this biased view — “Academic ‘Moneyball.'” The coverage is filled with the normal pros/cons treatment of opposing views, which creates false equivalencies and reduces critical thinking by making sure two jars of opposition are viewed as holding an equal number of beans, when one clearly holds fewer.
More alarming, the InsideHigherEd article spends more time outlining the positive arguments the authors make and only a few lines noting possible limitations. It turns out that authors and their mouthpieces both like positive results. But the study was small, the researchers’ route to defining the study group was questionable and highly manipulated, and there was nothing in place to probe the role of human judgment or non-data-captured facts in the outcome. It was, in fact, a very weak study with a malleable null hypothesis, if any. The study also was not “new,” as claimed by InsideHigherEd, but was published more than a year ago. During that time, it has been cited once.
What about providing a “balanced debate” rather than the boosterism seen here? That likely would have fomented ignorance, not knowledge, largely because such treatment creates more opportunities for coverage to fit into any existing biases readers may bring. As Robert Proctor, an expert in the spread of ignorance or “agnotology,” states it in a story from the BBC:
. . . ignorance can often be propagated under the guise of balanced debate. For example, the common idea that there will always be two opposing views does not always result in a rational conclusion. This was behind how tobacco firms used science to make their products look harmless, and is used today by climate change deniers to argue against the scientific evidence. This ‘balance routine’ has allowed the cigarette men, or climate deniers today, to claim that there are two sides to every story, that ‘experts disagree’ – creating a false picture of the truth, hence ignorance.
The “balance routine” seems a creation of the Internet Age, when space constraints no longer made it a priority to get to the point. “Balance” allows misinformation to feed confirmation bias. Examining this particular study of bibliometrics and tenure decisions more closely, you find problems the hundreds of Twitter sharers and dozens of commenters seem to have missed in their haste to indulge the confirmation bias Silicon Valley information economics has thrived on.
This is perhaps the ultimate damage wrought by the existing Silicon Valley information ethos — the belief that more information somehow compensates for worse information, or that at least there’s no real downside since we can see “both sides.” Clearly, that creates misleading equivalencies that do not help us understand the world.
The human mind is fallible, and full of biases that faster processing bring to the fore, allowing confirmation bias, hindsight bias, anchoring, and other biases to dominate our thinking. But we’ve fallen for it, and need to recover. This will require deep thinking and soul searching. As Stone writes in the BusinessWeek essay I quoted earlier:
For years, it was fashionable, even necessary, for tech companies to be an agnostic conduit for all flavors of web traffic. . . . It’s now apparent, in demands they refuse to build a Muslim registry, for example, that the old neutrality will no longer fly. This is the price of primacy: Silicon Valley is going to have to take a stand.
What stand will our industry take against fake news, superficial information purveyance, velocity-as-censorship, participation in a Muslim registry, attacks on scientific funding, constraints on academic or personal freedoms, or partisan attacks on scientific findings? Each scenario will likely have its own answer, but the days of simply shoveling more articles online without any thought of how they might be used, misused, or abused should be placed behind us. However, so far, our responses to predatory publishers, fake studies, fake scientific conferences, and fake peer reviews have been anemic and ineffective.
If there is one lesson from the past decade, it might be that more information feeds existing biases and can be exploited for nefarious purposes, all while turning a profit. The days of “neutral platform” disruption are behind us. It’s clear that some political operatives have found ways to exploit the Silicon Valley model and how it drives our thoughtless biases. It’s clear that simply putting out more information is insufficient to creating more knowledge. In fact, more information coming at us faster may do just the opposite. It may feed the bias beast by operating as de facto propaganda.
Part of the fix will involve becoming more skeptical and demanding of the information we tacitly endorse via publication decisions.
- Are those references in that submission truly relevant, high-quality, and meaningful? Or were they scraped during a desperate hour using Google Scholar to find a dozen sources to make the paper seem weightier than it is?
- Can the authors tell you what each referenced paper’s strengths, methodologies, and limitations are?
- Is the study design sufficient and robust, or is it full of holes, leaps of logic, and glossed over steps?
- How will audiences filled with confirmation biases treat a weak study that feeds a worldview?
- Will this detract from knowledge by creating a false equivalency?
Ultimately, change will require a change in economics and incentives. We continue to see payments and business models based on volume and scale of content, audience, and sizzle. We are now paid to put studies online.
Our measures need to be thought about. Social uptake isn’t usage. Usage isn’t always intellectual engagement. Traffic is not always benign. If purchasers were to force publishers to prove the quality of their offerings — interrogating their editorial processes, their reviewer pools, their retraction and correction policies, and their rejection rates — we may see a change in emphasis.
Markets respond to incentives, and as long as our incentives are Silicon Valley incentives — more information, faster, neutral to its political or social implications, falsely equivalent — we will suffer from the same maladies as the overall information space — biases, superficial thinking, delusional self-satisfaction. The disruption of our skepticism, logic, understanding of the world, and carefully conducted scientific research may be more than we should tolerate.