Undermining the logic of violence

Today nine countries are presumed to possess nuclear weapons; many more possess the capability to manufacture other classes of weapons of mass destruction (WMDs — biological, chemical and radiological weapons). This existential threat — and humanity’s innate innovative drive, plus perverse incentivization structures related to the armaments industry — means that we are racing towards ever more threatening capabilities. Still more countries are earnestly developing the capability to produce nuclear weapons within their territories. Concerns abound of non-state actors somehow procuring such a weapon.

An asymmetry of power has increased as weapons technology has advanced — very few — even one person — can have the power to end millions of lives and destroy huge areas of public and private property. But just as technological development carries risks, it offers opportunities. Here are three ways consensus networks and decentralized web technologies could provide a just and reliable mechanism to control the distribution, operation and deployment of weapons in the 21st century.

Blockchains are often touted as a way to dramatically enhance the efficiency of global supply chains, which are comprised of a vast constellation of private and public actors coordinating to transport goods across multiple jurisdictions from point of origin to point of consumption. The ecosystem is suited for blockchain because of the reliance on intermediaries, the participation of independent entities subject to varying regulatory requirements (i.e. countries’ customs rules), and its global nature. The establishment of transparent, strict protocols for global transport and logistics could smooth the process, reduce costs and increase integrity.

In weapons supply chains the term “point of consumption” belies an ominous truth: consuming armaments often means deploying them, detonating or firing them towards some target —with intent to damage or kill. International law typically deems deployment of such munitions by sovereign states in contexts of conflict just and legal, which explains the sustained year on year figures for continued production of light and heavy arms (SIPRI) — though questions have been raised about whether licit arms have been deployed criminally, for example by the Saudi coalition in the Yemen civil war (UN News 2018).

More relevant, perhaps, are situations in which legally produced and shipped weapons are diverted in transit to some illicit actor, often terrorist or other non-state militarized groups such as organized criminal networks. That violating a weapon’s “end-user certificate” is against the law (Bromley and Griffiths 2010) is often outweighed by the monetary or ideological incentives presented to intermediary (or even end-user) custodians of the weapons, explaining their decision to violate the certificate.

It is now possible to create a system of networked secure shipping containers, connected with smart contracts, to track the location and security of weapons and dual use goods as they traverse the global logistics network. This system would provide an additional technical mechanism to ensure compliance with end-use certificates, export control regimes and any other regulatory schemes designed to ensure the safe transport of dangerous materials to its intended recipient.

As an example, imagine a locked case of automatic rifles or centrifuges able to enrich uranium. The case is installed with sensors and connected to the Internet. The physical lock can only be “legally” (in the objective sense — upon receiving a signal from a smart contract) opened upon delivery to the pre-approved end user — that is, once the case arrives at its authorized destination, and the digital signature of the authorized recipient is received by the smart contract.

If the integrity of the case is compromised in transit, a notification could be transmitted to the network and relevant authorities could be alerted within minutes, along with information on who is the current legal custodian of the objects, its location history, etc. This system could include some dead man switch that would mean disruption of the data link would constitute an integrity breach just as a forced physical entry would. This system could be deployed in commercial contexts, but even could provide a system of oversight to the transfer of weapons in secure or military logistics networks. Provided that the keys are managed properly (i.e. malicious agents could not extract private keys, and therefore sign transactions that appear to be coming from the connected sensor), the system could provide an additional layer of security to the supply chains of weapons — of any controlled materials, really.

Extending this, in a secure enclave, secure multiparty computational geometry (Atallah and Du 2001) could enable privacy-preserving systems to detect errant actors without revealing strategic location data about compliant players.

Imagine a treaty governing where certain weapons could be stationed. Signatories — adversarial militaries — could submit encrypted location data about the positions of all of their weapons; within a secure enclave these positions could be verified to be in compliance with the treaty terms without revealing strategic secrets to any human actor. Only upon treaty violation would the misbehavior be revealed and agreed-upon punishments enforced.

The technological development of unmanned aerial, terrestrial and marine vehicles has inevitably led to the weaponization of such vehicles. These machines often rely on a consistent data link with human operators, but are increasingly using on-board software to navigate and make decisions. Debates regarding the encroaching reality that such on-board algorithms may be granted the ability to decide whether or not to engage a target — kill someone or destroy something — range from calls for an outright ban to claims that existing laws governing war will suffice in covering emerging and unforeseen combat technologies (Economist 2019).

Acknowledging that enforcing compliance with motivated actors intent on waging war will not be possible, I still wonder: might consensus networks and other decentralized technologies provide some way to ensure that autonomous weapons systems are governed by mutually agreed rules?

The concept here relies on specifications defined by participating militaries, perhaps contributed to by the public, that define the acceptable terms of engagement for an autonomous weapon system. Consensus networks would enable current, approved compiled programs (which could serve as a core program to run on an autonomous vehicle) to be maintained and updated by a community of developers. This solution, again, requires on trusted hardware: upon start, the vehicles would need to confirm with the network that they were running a current and unaltered version of software. This would enable the controllers of the system (a whole other conversation — perhaps experts and developers from cooperating militaries, say) a way to know that all the autonomous weapons they have deployed are operating on appropriate software. Autonomous vehicles could be programmed in a way that they could not operate with old, insecure, unapproved software.

This system has implications for all networks of autonomous vehicles — cars, ships, etc — but holds special relevance to ones that have the capability to deploy arms. As acknowledged, amoral adversaries would need serious incentives to comply with such a regime; perhaps making deployment of autonomous weapons that aren’t participating the system a war crime, or cause to implement economic sanctions, would be a start. Other incentivization schemes should be explored as the system would only be as good as the proportion of autonomous weapons participating.

Within the spectrum of destructive capabilities, some weapons stand out in their capacity to cause harm. Nuclear, biological and chemical weapons — collectively known as “weapons of mass destruction” (WMDs) are subject to international laws that other traditional armaments are not. Much effort is dedicated towards ensuring that states that do not possess nuclear weapons do not develop the capability to produce them (UN Security Council Resolution 1540). Control and governance of such weapons are among the most important responsibilities adopted by states that possess them.

Innovation resulting from developments in information technology may offer improved ways to govern the storage, maintenance and — most importantly — deployment of WMDs. Here we focus on perhaps the simplest application of distributed ledger technologies to governing the deployment of these weapons: the use of a multisignature smart contract to trigger launch.

This would require a privacy-preserving blockchain as part of the digital infrastructure of a nuclear-capable government, coupled with trusted computing environments installed on weapons and launch systems. Such a system — with nodes maintained by different branches or departments of the government, and possibly even allied governments or non-aligned third parties — could provide an immutable and highly reliable software system within which highly sensitive launch commands could be programmed to rely on several top decision-makers’ inputs. This could be extended to rely on signatures from outside a sovereign government, perhaps requiring an allied military to consent. If target coordinates were included into a smart contract call, further conditions could be programmed in dependent on the targeted area. It could conceivably be rendered objectively impossible (contingent upon the reliability of the trusted hardware) to fire an on-chain nuclear missile at an ally’s territory, if such a system were installed and configured appropriately. And of course it is worth saying this would only be a stopgap measure to complete denuclearization — a topic for another post perhaps.

Extending the concept of smart contracts underlying critical government and military digital infrastructure, certain executive powers could be enabled, say, only in the event of a declaration of war by an elected body — funds could be released, conditions for launch could be changed (i.e. making it easier for fewer decision makers to react to a threat) and so on.

The important question here: how might the new decentralized paradigm provide opportunity for innovation in how we govern our most harmful and powerful capabilities? Might this technology provide additional layers of security to prevent an unconscionable mistake or escalation? How could it be used to get ahead of a trend towards increasing automation of kill decisions? We have created computers to help us — how they might help us avoid the most grievous of human errors? Clearly the technology is too new to deploy in such a high stakes context now, but the brainstorming of today becomes the logical solution of tomorrow.

Undermining the logic of violence

Today nine countries are presumed to possess nuclear weapons; many more possess the capability to manufacture other classes of weapons of mass destruction (WMDs — biological, chemical and radiological weapons). This existential threat — and humanity’s innate innovative drive, plus perverse incentivization structures related to the armaments industry — means that we are racing towards ever more threatening capabilities. Still more countries are earnestly developing the capability to produce nuclear weapons within their territories. Concerns abound of non-state actors somehow procuring such a weapon.

An asymmetry of power has increased as weapons technology has advanced — very few — even one person — can have the power to end millions of lives and destroy huge areas of public and private property. But just as technological development carries risks, it offers opportunities. Here are three ways consensus networks and decentralized web technologies could provide a just and reliable mechanism to control the distribution, operation and deployment of weapons in the 21st century.

Blockchains are often touted as a way to dramatically enhance the efficiency of global supply chains, which are comprised of a vast constellation of private and public actors coordinating to transport goods across multiple jurisdictions from point of origin to point of consumption. The ecosystem is suited for blockchain because of the reliance on intermediaries, the participation of independent entities subject to varying regulatory requirements (i.e. countries’ customs rules), and its global nature. The establishment of transparent, strict protocols for global transport and logistics could smooth the process, reduce costs and increase integrity.

In weapons supply chains the term “point of consumption” belies an ominous truth: consuming armaments often means deploying them, detonating or firing them towards some target —with intent to damage or kill. International law typically deems deployment of such munitions by sovereign states in contexts of conflict just and legal, which explains the sustained year on year figures for continued production of light and heavy arms (SIPRI) — though questions have been raised about whether licit arms have been deployed criminally, for example by the Saudi coalition in the Yemen civil war (UN News 2018).

More relevant, perhaps, are situations in which legally produced and shipped weapons are diverted in transit to some illicit actor, often terrorist or other non-state militarized groups such as organized criminal networks. That violating a weapon’s “end-user certificate” is against the law (Bromley and Griffiths 2010) is often outweighed by the monetary or ideological incentives presented to intermediary (or even end-user) custodians of the weapons, explaining their decision to violate the certificate.

It is now possible to create a system of networked secure shipping containers, connected with smart contracts, to track the location and security of weapons and dual use goods as they traverse the global logistics network. This system would provide an additional technical mechanism to ensure compliance with end-use certificates, export control regimes and any other regulatory schemes designed to ensure the safe transport of dangerous materials to its intended recipient.

As an example, imagine a locked case of automatic rifles or centrifuges able to enrich uranium. The case is installed with sensors and connected to the Internet. The physical lock can only be “legally” (in the objective sense — upon receiving a signal from a smart contract) opened upon delivery to the pre-approved end user — that is, once the case arrives at its authorized destination, and the digital signature of the authorized recipient is received by the smart contract.

If the integrity of the case is compromised in transit, a notification could be transmitted to the network and relevant authorities could be alerted within minutes, along with information on who is the current legal custodian of the objects, its location history, etc. This system could include some dead man switch that would mean disruption of the data link would constitute an integrity breach just as a forced physical entry would. This system could be deployed in commercial contexts, but even could provide a system of oversight to the transfer of weapons in secure or military logistics networks. Provided that the keys are managed properly (i.e. malicious agents could not extract private keys, and therefore sign transactions that appear to be coming from the connected sensor), the system could provide an additional layer of security to the supply chains of weapons — of any controlled materials, really.

Extending this, in a secure enclave, secure multiparty computational geometry (Atallah and Du 2001) could enable privacy-preserving systems to detect errant actors without revealing strategic location data about compliant players.

Imagine a treaty governing where certain weapons could be stationed. Signatories — adversarial militaries — could submit encrypted location data about the positions of all of their weapons; within a secure enclave these positions could be verified to be in compliance with the treaty terms without revealing strategic secrets to any human actor. Only upon treaty violation would the misbehavior be revealed and agreed-upon punishments enforced.

The technological development of unmanned aerial, terrestrial and marine vehicles has inevitably led to the weaponization of such vehicles. These machines often rely on a consistent data link with human operators, but are increasingly using on-board software to navigate and make decisions. Debates regarding the encroaching reality that such on-board algorithms may be granted the ability to decide whether or not to engage a target — kill someone or destroy something — range from calls for an outright ban to claims that existing laws governing war will suffice in covering emerging and unforeseen combat technologies (Economist 2019).

Acknowledging that enforcing compliance with motivated actors intent on waging war will not be possible, I still wonder: might consensus networks and other decentralized technologies provide some way to ensure that autonomous weapons systems are governed by mutually agreed rules?

The concept here relies on specifications defined by participating militaries, perhaps contributed to by the public, that define the acceptable terms of engagement for an autonomous weapon system. Consensus networks would enable current, approved compiled programs (which could serve as a core program to run on an autonomous vehicle) to be maintained and updated by a community of developers. This solution, again, requires on trusted hardware: upon start, the vehicles would need to confirm with the network that they were running a current and unaltered version of software. This would enable the controllers of the system (a whole other conversation — perhaps experts and developers from cooperating militaries, say) a way to know that all the autonomous weapons they have deployed are operating on appropriate software. Autonomous vehicles could be programmed in a way that they could not operate with old, insecure, unapproved software.

This system has implications for all networks of autonomous vehicles — cars, ships, etc — but holds special relevance to ones that have the capability to deploy arms. As acknowledged, amoral adversaries would need serious incentives to comply with such a regime; perhaps making deployment of autonomous weapons that aren’t participating the system a war crime, or cause to implement economic sanctions, would be a start. Other incentivization schemes should be explored as the system would only be as good as the proportion of autonomous weapons participating.

Within the spectrum of destructive capabilities, some weapons stand out in their capacity to cause harm. Nuclear, biological and chemical weapons — collectively known as “weapons of mass destruction” (WMDs) are subject to international laws that other traditional armaments are not. Much effort is dedicated towards ensuring that states that do not possess nuclear weapons do not develop the capability to produce them (UN Security Council Resolution 1540). Control and governance of such weapons are among the most important responsibilities adopted by states that possess them.

Innovation resulting from developments in information technology may offer improved ways to govern the storage, maintenance and — most importantly — deployment of WMDs. Here we focus on perhaps the simplest application of distributed ledger technologies to governing the deployment of these weapons: the use of a multisignature smart contract to trigger launch.

This would require a privacy-preserving blockchain as part of the digital infrastructure of a nuclear-capable government, coupled with trusted computing environments installed on weapons and launch systems. Such a system — with nodes maintained by different branches or departments of the government, and possibly even allied governments or non-aligned third parties — could provide an immutable and highly reliable software system within which highly sensitive launch commands could be programmed to rely on several top decision-makers’ inputs. This could be extended to rely on signatures from outside a sovereign government, perhaps requiring an allied military to consent. If target coordinates were included into a smart contract call, further conditions could be programmed in dependent on the targeted area. It could conceivably be rendered objectively impossible (contingent upon the reliability of the trusted hardware) to fire an on-chain nuclear missile at an ally’s territory, if such a system were installed and configured appropriately. And of course it is worth saying this would only be a stopgap measure to complete denuclearization — a topic for another post perhaps.

Extending the concept of smart contracts underlying critical government and military digital infrastructure, certain executive powers could be enabled, say, only in the event of a declaration of war by an elected body — funds could be released, conditions for launch could be changed (i.e. making it easier for fewer decision makers to react to a threat) and so on.

The important question here: how might the new decentralized paradigm provide opportunity for innovation in how we govern our most harmful and powerful capabilities? Might this technology provide additional layers of security to prevent an unconscionable mistake or escalation? How could it be used to get ahead of a trend towards increasing automation of kill decisions? We have created computers to help us — how they might help us avoid the most grievous of human errors? Clearly the technology is too new to deploy in such a high stakes context now, but the brainstorming of today becomes the logical solution of tomorrow.