When DAOs are Weaponized

Vulnerability

In Nick Bostrom’s Vulnerable World Hypothesis, he describes scientific and technological progress as a force that may lead to an increase in the vulnerability of the world by decentralizing access to world-destabilizing technology and information. It is by luck, according to Bostrom, that we have not yet been able to produce a technology that leads to the destruction of civilization. Consider the impact if we were to have discovered that the process of nuclear fission could be facilitated with simple materials found in the average household. That knowledge would invariably facilitate the destruction and destabilization of society through countless detonations of DIY atom bombs.

Or consider the impending technology that decentralizes the ability to print genetic material, and enables those with only an undergraduate-level understanding of biology to download open-source genomes and synthesize viruses at home. With a low barrier to entry, such a technology will all but guarantee the deaths of millions in engineered pandemics.

Notably, it is not the power, per se, of the technology that leads to its inevitable weaponization, it is the decentralization and distribution of such a technology. Decentralization makes regulation impractical, and distribution helps ensure that at least one ignorant, incompetent, or malicious person will wield that power to the great detriment of human civilization.

I propose that some decentralized collaboration technologies such as blockchains and DAOs may be an understudied and unanticipated long-term source of similar vulnerability to the world.

The Stability of the Status Quo

I once watched a discussion in the Senate in which an amendment to a new bill was proposed (The Lummis-Wyden-Toomey amendment to the 2021 Infrastructure Bill). The amendment required and attained unanimous support from all senators, but did not pass because a single senator announced he would vote against the amendment unless everyone also agreed to an entirely unrelated amendment. Of course, others disagreed with the additional amendment, and it therefore blocked the original amendment from passing.

At the time, I lamented how broken our system is in its ability to effect change. How could it be that such “pork barrel politics” could render an amendment with literally unanimous support unpassable?

But I’ve since had a terrifying thought: What if it’s precisely the fact that our systems of collaboration yield political parties, incumbent advantage, and tyranny of the status quo that gives society the stability it needs to continue its slow, safe, banal procession into the future?

Impractical collaboration

Imagine that you’re an extreme antinatalist whose goal is to forcibly sterilize all humans, or you’re a religious extremist with a mission to annihilate all apostates. You’re in a tiny minority of people worldwide, and up until recently, you faced two significant challenges to realizing your goals: recruiting others and organizing your efforts.

Recruitment

The challenge of finding others with whom to collaborate is a difficult one. Outing yourself as a dangerous extremist carries reputational risk and it opens the possibility that others will detect and interrupt your efforts. It’s often unlikely for those with extreme views to find others in their local community with similarly extreme views, therefore successful recruiters must also be technologically savvy enough to anonymously navigate the internet and avoid detection by authorities.

Organization

Even after successfully establishing a group of like-minded people, there are many other difficult challenges: how to facilitate ongoing anonymous interactions, how to evaluate newcomers and manage the organization, how to buy the needed materials, where to store physical goods, and how to move money around. Traditionally, the only coordination structures available are traditional corporate or hierarchical structures that expect a top-down approach to organizing the group’s efforts. In addition to the difficulty with the organization of any nefarious activity, this comes with all the standard issues with any centrally-planned effort:

• People are forced to work on tasks they don’t find personally compelling, decreasing individual efficiency.
• Specified tasks are rarely the optimal path toward realizing the goal, decreasing collective efficiency.
• Central points of failure pose existential threats to the life of the organization, for example, if those at the top of the organization leave for any reason.

But none of these issues are endemic to the general problem of coordination and instead are derived from the centralized coordination structure. There has never been a coordinated effort to produce extremely efficient decentralized organization tools… until now.

Decentralized organization is coming

Many see the internet as the substrate beneath all future human collaboration efforts. This is sometimes known as the Network State hypothesis. Driven by this hypothesis and also by the blockchain industry generally, there’s an ongoing “arms race” to optimize the algorithms and their parameters used to coordinate community efforts. Many Decentralized Autonomous Organizations (DAOs) are simply traditional centralized power structures that use the blockchain as an automation and security technology to better manage internal affairs. These DAOs might be more efficient than their traditional non-automated corporate structure counterparts, but they do little to advance the underlying mechanisms used to collaborate. But there are other ongoing experiments which are innovating on the basic tools for decentralized coordination. These are an attempt to create what we could call Optimized Decentralized Organizations (ODOs).

Self-organizing using ODOs will allow decentralized communities to realize the benefits of economies of scale to compete with efforts that previously would have required top-down central planning. Eventually, tools will exist that operate on entirely anonymous, peer-to-peer networks. Moving money around will be trivial and private. There will be no central point of failure in the organization. The ability to coordinate effectively will itself have become decentralized and distributed.

Radicalization

The Law of Group Polarization was an observation by Cass Sunstein that like-minded groups tend to self-radicalize or drift towards ever more extreme versions of their shared mindset. Based on our limited prescriptive understanding of this radicalization process, we have no reason to assume that groups using decentralized collaboration techniques will be immune from its effects. Where today factions within a larger group stalemate each other’s progress, in the future each faction may be able to effectively mobilize their supporters to make significant progress on their contradictory goals. Each group will tend to adopt an increasingly extreme version of their ideals, further polarizing society. In the worst cases, groups will sufficiently self-radicalize to an extent that leads to an increase in violence and injustice.

If it only takes a single maniac to effect massive damage to society through a suicide bombing, or the release of a synthetic virus, imagine the magnitude of the effect of a single, global, anonymous, decentralized, coordinated effort to intentionally wreak havoc on our fragile global infrastructure.

We, humanity, will either deliver on this promise of unlocking the potential for previously uncoordinated groups to effectively coordinate, or we will not. If we do, then an increasing number of ideologically radicalized groups will have their ability to effectively coordinate enhanced. If we do not, then we are left with the banality of the status quo and the fragile inefficiencies of gradually centralizing organizations.

What can be done?

In order to combat such existential threats, Nick Bostrom has posited that we would need a system of ubiquitous real-time worldwide surveillance, which he has called, “turnkey totalitarianism.” From my perspective, any effort used to achieve turnkey totalitarianism would introduce a central point of failure extremely susceptible to attack, corruption, or misuse. Therefore, I propose nearly the opposite approach: radically decentralize our global infrastructure. Supply chains, energy grids, internet nodes, and many other systems fundamental to modern society are extremely vulnerable to targeted attacks. Eliminating all central points of failure in our infrastructure would go a long way to improving the resiliency and anti-fragility of society. I acknowledge, however, that this doesn’t prevent disaster, but simply reduces the impact of many failure modes.

Conclusion

Assuming we fully realize the promise of ODOs, groups that are not harmful or extremist in nature will also be able to collaborate. This will undoubtedly bring huge benefits to society, encouraging technological, scientific, and societal advances. In fact, it’s the promise of these benefits that motivates the techno-optimists of the world to continue their innovation faster than anyone can understand its implications.

The situation is reminiscent of the state of research into artificial intelligence, which has long been described as a technology with the potential to bring both unparalleled good and harm to society. And just as the efforts to develop general artificial intelligence systems are decentralized and therefore inevitable, so too are the efforts to realize Optimized Decentralized Organization techniques. It is also highly unlikely that either industry or any truly decentralized effort will be able to be effectively controlled through regulation. There is therefore nothing to be gained by efforts to stop innovation in this area. Such efforts will definitionally be most effective to those most concerned about the issue, who are precisely the people we would want at the forefront of the relevant innovation. Rather, we can only proceed with an increasing awareness of the risks posed by such innovation, and through our collective ingenuity, work to thread the needle and safely deliver society into a prosperous future.