Jan J. Zygmuntowski has been our first Open Future Fellow, working on issues related to public data commons. As a contribution to his fellowship, we’re publishing his opinion on the Paradox of Open, which builds – and responds to – the series of Paradox of Open Responses that we recently published.
It is heartening to see a debate erupting not on the usual bread & butter of policy, but on key principles and strategies to make ideas come to life. This is what Alek Tarkowski and Paul Keller achieved with their provocative Paradox of Open, and the amazing responses that followed. What particularly stood out to me was James Boyle’s critique Misunderestimating openness, both because it was witty and because it presented a defense of the open against the discontent. But I also disagree with it fiercely, and so I decided to write this follow-up, a response to a response. It is well advised to start with the previous texts.
Additionally, I offer here a seemingly contradictory resolution of the Paradox of Open because, obviously, you fight paradoxical fire with fire! The open, I claim, requires gatekeeping.
But let’s start from the beginning.
It is easy to disregard the technolibertarian promise of the early Internet days, best epitomized by Barlow’s “Declaration of the Independence of Cyberspace.” After all that followed – the expansion of the digital to every corner of economic and social life and the rise of powerful, global “cloud empires” – we tend to forget how radically different were the visions sold to us mere 30 years ago. But it was indeed a con that we desperately wanted to hear in a world of defeated utopianism, a con repeated many times later in a similar manner. From the dot-com bubble, through Web 2.0 and the sharing economy, all the way to Web3 and crypto. Of course, the blame is not on the open movement only (or even not at all!). But if calling the broken promise out is a strawman argument, then you could litter half the planet’s rich farmland with strawmen, each bearing a specific name straight from elite conference halls.
Keller and Tarkowski interrogate openness as a feature leading to social progress precisely in this context. Nobody ever argued for any concept as a good in itself, detached from its benefits and development of our society. Take the other examples from Misunderestimating: free speech wasn’t recognized as a thing until recently, and the popular franchise is an achievement regained after centuries of feudalism, slavery and colonial capitalism. These ideas won because they offered a vision of a better society – one less riddled with antagonisms, creating new opportunities and better fulfilling the human desire for justice. If open is not about – in the words of Keller and Tarkowski – “a more just and egalitarian digital society” – then what’s the point?
Despite Boyle’s attempt to remain free of judgment, he does provide us with a hint of a particular vision. “It [a bet on openness] is a belief that, in the long run, open systems are better for the human race.” What exactly does “better” stand for, if not for the fulfillment of certain values like, say, egalitarianism or justice? Forgetting that “better” has to convey a particular value amounts to losing track of why anyone bets on open in the first place.
On many occasions, Misunderestimating claims that everybody knew “openness could lead to bad outcomes as well as good.” This claim is not without doubt, and one could easily argue that neglecting distributive consequences was a larger symptom of the eradication of egalitarian thinking, political economy and philosophy post-1989, in the period of collective hallucination of the “end of history.” But let’s take it at face value and assume everybody knew, meaning the bad outcomes of unrestricted openness fit the known-knowns quadrant of the Johari window (“we know that we know”).
One would then assume that such great conscious understanding leads to specific policies and actions. As a young scholar coming to this field in the wake of sharing economy buzz, I wondered what precautions were taken so that the open web is not used to accumulate capital and suck the wealth of local communities on a planetary scale. I also wondered how B2G, or rather business-to-public data sharing, is going, given that we broke government silos like piggy banks. Then I wondered if anyone made sure large platforms never bring copyright back, or implement closed protocols on the layer of interoperable ones. And frankly, it seems that leaving the supposed known-knowns idle created a disenfranchisement so intense that even monkey-toting crypto snake oil pushers seemed for many a viable alternative route for the web.
Ideas have consequences that play out in an unequal capitalist society, and this has to be taken into account instead of dismissed as not worthy of consideration. Boyle ridicules the naivete of those who wish to “improve” upon Bible translations, democracy, or free speech, as if it necessarily led to the erosion of these hard-won milestones of progress. One could read it as: “open must be open only, do not touch!”. The mission of open movement ends with openness; cleaning the resulting mess is up to everyone else, even if nobody was dumb enough to believe openness only comes with the perks. I call this open absolutism because it reminds me fondly of free speech absolutists, who contribute to society greatly when they stand up against censorship, but fall flat when actually running social media; they abdicate of any responsibility for protecting free speech because it requires prosecution of some types of speech. Let that sink in.
Instead of creating collective vehicles for the advancement of technology-backed welfare, the open movement acted like a double-edged sword. On the one hand, open advocates liberated us from the previous institutions of power, unfit for the emerging network society. On the other hand, they unintentionally paved the way for the largest platforms to establish their dominance and take aim at the open itself. The question Boyle poses a few times – whether oligopolists benefit more from open or proprietary resources – works only as an exercise done in a historical vacuum. It detracts from the issue at hand and doing things smarter than the previous generation of activists. Just one such example:
“Thus, it is very easy to blame openness. But the key question is the alternative. Do you think we would be better off with censorship, dictatorship and a proprietary world of information sources?”
Boyle tricks the reader into thinking this is an open question, but actually closes the space of possibility to just two. It’s either open or dictatorship. Smart move! One undoubtedly learned from Thatcher’s TINA, which forced society’s hand into playing neoliberal despite all the evidence that regulated markets outperform unregulated. In this context, I find interesting this observation:
“That technological, political economy was never going to be a stable equilibrium. If anyone believes it was permanent, Open is a wake-up call.”
To find that the economy is dynamic and ever-shifting instead of constantly falling back to some equilibrium is already proving smarter than the average econ 101 textbooks. But to assume that in 30 years since the birth of the internet, we will watch both the rise of open and its own undoing at the hands of previous allies is another thing. Institutional settings may last longer or shorter; defend the new, better status quo with various means. Yet some political economies fall apart forever, as slavery or child labor did. Or rather, they will not return, unless our institutions let bad people force them on us. So the choice is not “open or dictatorship,” but rather how to remain as open as possible in a dynamic technological and economic setting.
And here comes my central argument, and a solution of a kind to Keller and Tarkowski’s intellectual puzzle. What I claim is that the “Paradox of Open” is, in fact, a new telling of the old wisdom of the Paradox of Tolerance. A tolerant society, according to Karl Popper, cannot tolerate the intolerant – those who jeopardize the existence of a tolerant society in the first place. Respectively, open cannot tolerate privatized gains and new enclosures. But the heart of the Paradox lies in the antagonism. The tolerant have to stop the intolerant before they succeed; the open have to exclude those attempting enclosures. In the idealistic world of Plato, this is self-contradictory, but in the real world, such is the experience of anyone who knows the dangers of any absolutism. The open, I claim, requires gatekeeping.
While admitting that the commons are a blueprint for understanding the unaccounted-for economics of the public, shared, social and collective, Boyle observes that leading voices at the time believed there was no commons theory. Actually, not only did we already have a burgeoning commons theory, but it even made it to the Sverigesriksbank “Nobel” Prize with the acclamation of Elinor Ostrom’s groundbreaking work on common-pool resources. What I suspect is that the findings of commons theory did not fit what some of the open activists wanted to hear – hence neglecting it was preferred to adopting in praxis and breaking some dogma in the process.
So what can we learn from the commons theory, studied in the “wilderness” of fisheries, forests, pastures and water resources? That boundaries matter. That sustainable use prevents depletion. That access can be held relatively open but monitored and sanctioned for misuse at the same time. Local communities need degrees of freedom from the one universal policy that is formulated by distant decision-makers on the hill.
There is much discussion about to what extent the commons theory directly transposes to the intangible realm, especially given the near-zero replication costs of data. But various authors, including myself, argue that just like any other commons, digital can withstand rivalry only to some extent. After a certain point, the trust breaks, produsage declines, and/or the winner takes all and offers a new walled garden. Privacy is a rivalrous good, and so are the outcomes of using and producing with the commons. If data leaks, privacy is lost – no matter the costs of copying data across the web. Similarly, capitalizing on digital commons, like Google did on open source Android system, disincentivizes others to contribute to the open and instead promotes going corporate. Unguarded – or should I say “ungatekept” – the open fulfills the “tragedy of the commons” (unsustainable use and freeriding leads to deterioration), which real commons solve and are resilient to thanks to strong governance mechanisms.
Following Ostrom’s invitation to go “beyond state and market” means thinking beyond the distinctions between public and private, open and closed. The commons just don’t fit these categories. Therefore the question is not whether open is better than closed (duh) but for whom and on what terms should we open.
One could dismiss this idea of stewardship, calling it “political,” claiming it will be biased, captured or lagging behind technology. Well, maybe. But politics, or at least political economy, is here anyways.
The politics of open are one that favors disruptors and challengers, but fails to secure itself and the larger society from further monopolization, rent-seeking and private power.
When it comes to the war on copyright and the opening of science, I have nothing to add to Boyle’s sound argument. His comment on the right to repair is exceptionally telling of what a moderate position is. “We need a campaign to fight the anti-competitive abuse of intellectual property rights” is not a call for 100% open knowledge at all times. And that is absolutely okay! Moderation is well advised while tampering with complex innovation systems, since IPRs sometimes incentivize entrepreneurial, creative small guys. However, that same moderation can be granted to communities wishing to choose who and on what terms get to access their data. It is the really existing alternative to open absolutism.
And because Lessig was right, access to data supersedes legal claim to it. If that sounds iconoclastic to all the lawyers out there, think about what Big Tech did with our data and how, after GDPR and all the court losses, they just keep paying fines and chugging along. Hence the opening and access to data should be our sole focus now, and legal titles or rights only the latter. We have to build institutions and infrastructures capable of governing data commons, with all the technical means to exclude anyone who abuses our terms. Thus a predistribution of gains and precaution towards harm will be achieved. Because what cannot be done, we have no need to prosecute.
“Data is the new oil,” repeated by Boyle, is a worn and utterly wrong saying. No, data is not extracted from passive reservoirs; it is a product of human labor. It is not a commodity but a relational infrastructure. And it has to be stewarded like commons, because otherwise, those capable will be the first to rip what our collective intelligence sowed.
This leads us directly to the ongoing debate du jour, namely, training AI on open data sets. Boyle reaffirms his open faith with statements such as: “High-quality, ethical datasets, openly licensed, can increase competition, decrease inequality and promote transparency.” So, was that political economy lesson learned after all? It just sounds as if a data-grabbing startup asked ChatGPT to produce a mission statement. In fact, I did just that.
Prompt: Write a single-sentence mission statement saying why open ethical datasets are good.
ChatGPT: Open ethical datasets are good because they provide transparent, unbiased data that can be used to advance scientific research, improve decision-making, and benefit society.
We have to problematize this field finally, before jumping to such nice soundbites. What does it even mean for a dataset to be ethical, given it’s a trait of human character and choices? More than often, ethical is a code word for self-regulation to stop regulators in their tracks. Are we sure open datasets increase competition, if the capacity to train is so dependent on first-mover advantage in ML talent, computer architecture and funding to burn massive amounts of energy? For the vast majority of the world, the dependence on few proprietary AI models – trained on ethical, open data sets, sure! – is tantamount to technocolonialism. I would rather argue that ethical datasets are ones that purposefully curate high quality and open access according to terms set by society, and not just by anyone – as if political economy never mattered.
Gatekeeping our common data is the right thing to do if we wish to see AI trained for socially beneficial purposes. Example: our health data should be governed by our collective choices, and that requires both opening existing silos (public and private) and creating democratic institutions which review requests for access, control the access and make sure the gains are distributed accordingly. There are viable alternatives discussed right now, such as public data commons. Because allowing massive value leakage in hopes of getting a discount on tokens for corporate AI would be, in fact, disastrous for inequalities and the provisioning of public services.
The description of the open movement as the intersection of the advocates of the commons (Creative Commons, commons-based peer production, public domain) and the open web movement (decentralized global architecture, net neutrality) is very useful. This alliance has worked miracles, effectively creating a geographically-agnostic, open web full of ideas, knowledge, shitposts, art and your daily sludge of breakfast content.
However, what if that intersection was only possible in a given structural setting, where the powers-that-be of telecoms and copyright owners shared one goal: “strangling the open web at birth” and keeping respective market dominance? This is not to say that since we won the old battles, the borders are fixed forever. Just think of copyright filters on platforms or free Internet access to certain websites provided by telecoms (zero-rating) against the spirit of net neutrality. Without a doubt, “all that is solid melts into air,” including the open web we already have. But because the structural setting has changed – and so have threats in the digital – it is worth considering whether a new programmatic alliance and recalibration of positions are desirable.
Where does that take the open movement? Boyle is in favor of keeping it pure from all the other currents, like the new antitrust wave or the surge of decentralization. But to remain in the same position of open absolutism is to admit all went right. What a luxury! The privacy folks, although winning many fights both in parliaments and courts, seem less self-conceited when they denounce the perfection of consent and contractual thinking in favor of social harms, justice and collective rights.
Consequently, one can belong to the open camp and still seek out new means of doing the open. Boyle stresses rightly that “property rules – whether open or closed – are singularly poor privacy tools.” Let’s read it as a recognition that open movement should pick the right tools to win, party, and have less painful hangovers when the consequences unfold. Because open absolutism is a noble cause bringing a knife to a gunfight.
Is my contribution just the usual manifestation of progressive Zeitgeist or rather a Molotov thrown in the revolution of the digital? I agree with Boyle that nobody knows for sure if the popular opinion is right, but to disregard so many voices seems to be quite a close-minded thing to do. We often laugh at fancy tech, looking for a problem to solve. But if the open movement fails to take into account what the popular sentiment is, and what it asks for, it will become just as detached. And my guess is none of us – Alek, Paul, James and myself – are fond of such a scenario. I am just a newcomer here, but see the tremendous strength of the open in its own openness to critique and to other movements. So, just as once it embraced the contenders of the digital economy, it can very well now recalibrate its positions and aid all the experience and good faith to new allies.