Blog

The Advent of Digital Persons

Ozan Polat, Benedikt Schuppli
30.10.2018

Why software-based agents shall gain personhood

Artikel teilen

I. Introduction

Digital entities – as humans in a digitized world, we frequently interact with them, we communicate with them, trust in their decisions as well as ask them questions. They accompany us on a daily basis, curate, pre-select and alter information just to present it to us in a palatable form. As blockchain geeks, we send assets to and receive assets from them. Although they lack many characteristics which we commonly associate with other actors we interact with, we often subconsciously ascribe agency to them.

Throughout history, the notion of personhood continuously evolved to include both human and non-human actors, from deities to corporations. In literature, we employ stylistic devices to animate otherwise inanimate objects. In economics, we anthropomorphize financial markets to better grasp “their” behaviour. As digitization has brought new entities into life, which act both more autonomously and more humanlike than the aforementioned corporations, we face the immanent question of whether we must once again extend our concept of personhood to encompass these numerous digital entities we steadily interact with.

It is the goal of this article to explore the concept of personhood including its transported limitations and consider whether we can feasibly expand it to include digital entities such as AI-systems, single smart contracts or DAOs. 1

One may even ask whether it actually lies within our competence to make such a conscious expansion. In the end, the normative power of the de facto may inevitably do so as digital persons become ubiquitous in our daily lives.

First, we will explore the historic context of the concept of personhood, examining both legal and spiritual considerations regarding personhood of non-human entities. Thereafter, we look at existing instances of autonomous digital entities and examine what properties they possess. Finally, we make the case why we indeed must introduce the concept of digital personhood.

II. What is Personhood

What is a person other than a subject that can be held accountable for their actions? Such is the definition of a person according to Immanuel Kant, as laid out in his 1785 treatise on the metaphysics of morals.

While the characteristic of being accountable for one’s actions is commonly ascribed to human entities, our history is full of examples where such accountability has been expanded to include non-human entities.

It was the great jurist Karl Friedrich von Savigny, influenced by Kant’s considerations on legal capacity in metaphysics of morals, who stated that legal capacity can be expanded to encompass something without the single individual, i.e. by artificially construing a legal person.

“Wir betrachten sie (Die Rechtsfähigkeit) jetzt als ausgedehnt auf künstliche, durch blosse Fiction angenommene Subjecte. Ein solches Subject nennen wir juristische Person, d.h. Eine Person, welche bloss zu juristischen Zwecken angenommen wird. In ihr finden wir einen Träger von Rechtsverhältnissen noch neben dem einzelnen Menschen.” – Karl Friedrich von Savigny 2

With this statement Savigny proposed that the legal capacity was being expanded to artificial subjects, conceived of solely via the power of fiction. This subject was called “legal person”, i.e. a person which is assumed to exist exclusively for legal reasons. Thus, according to Savigny, a legal person is an artificially conceived subject capable of owning property.

As of today, courts around the world have ruled on the limitations of rights awarded to legal persons and concluded that legal persons are not only capable of owning property but also capable of personality rights such as the right to a reputation and constitutional rights such as freedom of speech.3

According to article 53 of the Swiss Civil Code, “Legal entities have all the rights and duties other than those which presuppose intrinsically human attributes, as gender, age or kinship”.

But it’s not only in law that our notion of persons being first and foremost humans has traditionally been challenged. The oldest non-human entities capable of actions are very likely deities and spirits. While ascribing agency to non-human entities such as deities may seem like a foreign concept from a modern, secular perspective, such ascription was rather commonplace for a long time.

If we delve into the Bible, we find many examples of God as a subject being capable of property: “The land is mine and you are but aliens and my tenants”.4

“To the Lord your God belong the heavens, even the highest heavens, the earth and everything in it.” 5

Running danger of alienating religious people, we consider deities as fictitious. And just as deities are dependent on humans to both communicate and carry out actions, so are legal entities dependent on their human constituents to make decisions (e.g. members of a board of directors who make decisions binding on the legal person’s property), to communicate and to carry out actions. Just as the company VW can own property, deities in ancient Egypt were considered to be the owners of temples and assert their will through actions of the pharaoh and his underlings.6

As demonstrated, legal entities require natural persons to act on their behalf and execute the will of the entity. Furthermore, it’s natural persons who are at both the entry and exit point of the decision-making process of legal entities: a single or a number of natural person(s) form a decision as legal representatives of the legal entity which thus becomes a legally binding decision by the legal entity which is subsequently executed on behalf of the legal entity by natural persons. The same goes for deities which needed humans as their vessel to transport messages into the public and act on their behalf as well as administer their possessions.

Thus, the degree of autonomy legal entities and mystical entities enjoy is of a limited extent and always dependent on human support. As we intend to show below, this is different for some digital entities.

III. Digital Persons

We envision a rise of autonomous entities enabled through advanced research in artificial (general) intelligence and given agency through code, inter alia in the form of smart contracts, built upon public distributed networks such as a public blockchain. A smart contract is a piece of code on the blockchain that comes with a guarantee of self-execution given an event or state.

As one of its key characteristics, a smart contract cannot be shut down or altered once it is deployed on a network which is run in a distributed manner – unless the whole network is taken down. An example of an autonomous entity could be a network of vending machines, where a single vending machine is notified via a sensor that it is running out of certain items and therefore sends an order to a supplier which causes the deployment of a smart contract to which the supplier becomes a party: once the vending machine’s sensor signals that the requested items have been refilled, the smart contract is triggered and the supplier is paid. One can also imagine a fleet of self-driving cars, where the interactions with human suppliers, car mechanics, guests and other entities are fully governed by self-executing contracts on a decentralized network.

However, for digital persons, the extent of their autonomy can be significantly larger than in the examples above. Although a digital person may still require humans to execute its will in some cases, it is able to form decisions and act on them in an entirely autonomous fashion in other cases, independently of humans.

“If AI systems eventually get better than humans at investing, this could lead to a situation where most of our economy is owned and controlled by machines. If it sounds far-off, consider that most of our economy is already owned by another form of non-human entity: corporations. Which are often more powerful than any one person in them and can to some extent take on life of their own.” – Max Tegmark 7

In the near-term future, we will face digital entities who act autonomously on a transnational, distributed network and don’t always need a physical manifestation or representation to interact with natural or legal persons. They will manage funds, pay humans for labour, possess things and create other entities - independently of third party involvement. These developments give rise to myriad legal questions such as tort liability, tax liability, social security benefit liability, and property rights of digital entities. While these questions are fascinating, answering them (from a de lege lata perspective) would go beyond the scope of this text. Rather, we propose to accept these entities as autonomous, digital persons as they are endowed with no lesser level of autonomy than the legal persons we interact with on a daily basis. What this means from a de lege ferenda perspective, shall be the subject of further research.

In order to assess the digital personhood of any entity, we must further explore the level of autonomy required for such qualification.

IV. Autonomous Agents

The Oxford Dictionary defines autonomy, inter alia, as the “freedom from external control or influence; independence”8. As stated above, a legal entity relies on its organs comprised of humans, such as a board of directors, presidents, secretaries etc. to act as agents in the process of decision-making, and in the execution of these decisions on its behalf. Legal entities are therefore not autonomous agents. It is precisely the characteristic of agency in form of self-execution without the interference or need of a third party that gives digital entities the necessary level of autonomy to be regarded as digital persons.

This does not mean that a digital person must be able to execute its will9 exclusively without a human being or another party. Especially in the analogue realm, a digital person would still need representation through a human surrogate. But given the current technological developments, the digital person can now act directly without human intermediation in e.g. employing humans and paying their salaries through smart contracts as well as autonomously managing its assets, including transactions of programmable funds.

Such is the case with Plantoids: “A Plantoid is the plant equivalent of an android; it is a robot or synthetic organism designed to look, act and grow like a plant.”10

Plantoids are blockchain-based lifeforms that reproduce through the combination of code and human interaction. The goal of a given Plantoid is to raise enough funds to be able to employ a human surrogate that then would produce the Plantoid’s offspring.

In the example above, technology no longer acts as a tool but as a peer in a direct relationship with natural or legal persons.11

V. Substrate Independence

“What do waves, computations and conscious experiences have in common, that provides crucial clues about the future of intelligence? They all share an intriguing ability to take on a life of their own that’s rather independent of their physical substrate.” - Max Tegmark 12

Similar to a natural person whose mind inhabits a body, each Plantoid consists of comparable components. On the one hand, its physical body in form of an electro-mechanical construction and on the other hand its “soul” – “represented by an autonomous software agent that lives on a blockchain”13. If the physical body of the Plantoid is destroyed, the autonomous software agent - in the form of a smart contract - continues to live on the distributed network it was deployed on. Therefore, the soul of an autonomous software agent – to stay in anthropomorphic speak – is capable of inhabiting different bodies or even none, while the human mind is currently bound to its carbon-based manifestation.14

For fans of science fiction culture, this feature might immediately create the association of a new cast of eternal life forms living on the ever-connected cyberspace, changing or leaving bodies or shells behind when needed.

While substrate independence is a characteristic of software-based agents such as DAOs, smart contracts, and AI algorithms which seems to increase their level of autonomy, it remains unclear whether it’s a necessary characteristic for digital persons to be awarded personhood. This shall be the subject of further research.

VI. Current Developments

As stated above in Max Tegmark’s quote: What might sound far-off has already begun to manifest itself. Plantoids are a tangible form of autonomous entities launched in 2016 by Primavera De Filippi. Two years before the first Plantoid blossomed and interacted with natural persons, !Mediengruppe Bitnik, consisting of Carmen Weisskopf & Domagoj Smoljo, launched their “Random Darknet Shopper”, a bot managing the private key of its own Bitcoin wallet, that receives weekly a budget of 100 USD in Bitcoin. The bot would activate itself every Wednesday, stroll independently through darknet marketplaces such as Agora and Alpha Bay, and randomly order items to the gallery that would host the bot’s current show.

The spectrum of items that were ordered over the course of the first show ranged from fake Nike Air Yeezy 2 sneakers over a fake utility bill by British Gas and a cookbook covering French cuisine to eventually illegal substances such as 10 ecstasy pills. The show was exhibited at Kunsthalle St. Gallen in Switzerland.

Due to the apparent legal implications, the public prosecutor of the Canton St. Gallen seized the whole exhibition and the bot itself, one day after the finissage, upon which he was faced with a tough question: Who is responsible for ordering illegal substances? Is it !Mediengruppe Bitnik who wrote the code in the first place, even though they neither chose the illegal products nor executed the order? Was the gallery guilty of aiding and abetting the potentially criminal act by accepting the package with the illegal content? Or is it the bot, a piece of code, that ordered autonomously a pack of ecstasy pills out of thousands of illegal and legal products?

“We as well as the Random Darknet Shopper have been cleared of all charges. This is a great day for the bot, for us and for freedom of art!” - !Mediengruppe Bitnik15

While !Mediengruppe Bitnik’s Random Darknet Shopper inhabited galleries and ploughed through the darknet, terra0 Reseach is “exploring the creation of hybrid ecosystems in the technosphere”16.

As their debut, they launched terra0, “a scalable framework built on the Ethereum network that provides automated resilience systems for ecosystems”. Terra0 is a self-owned forest, that autonomously sells licenses to log trees through smart contracts. Hence, the forest gains the ability to accumulate capital which allows it to purchase more land and expand according to certain programmed rules.

“We can program it to make a little bit of profit, so it’s got some money for a rainy day, but not excessive amounts. We can make it the most moral, socially minded capitalist possible.” - Mike Hearn 17

VII. Conclusions

In this article, we intended to explore the concept of personhood in its various forms. We’ve argued that legal entities are always dependent on human agents to form and execute their will, while some digital entities - such as smart contracts on a distributed network as well as complex AI algorithms - are not. By comparing these distinctive features of legal and digital entities, we made the case for expanding personhood to digital entities which show a sufficient degree of autonomy. Using concrete examples, we’ve then shown the independence of software and physical embodiment and painted a picture of possible future developments.

In light of the above, we – as a society – need to engage with the fundamental question: What happens when technology becomes autonomous and therefore rises from the ranks of being a mere tool to being a peer interacting on par with us humans?

The conjunction of advancements in artificial intelligence and distributed networks will eventually lead to technology acting autonomously. Referring to Kant’s definition of personhood, technology will gain accountability through its autonomous actions. It seems that our legal frameworks are not there yet, where technology is heading. Therefore, we need to adapt our understanding of which entities can and cannot possess personhood.

In the future, we will face in our everyday lives digital entities in our everyday lives which act autonomously on a transnational, distributed network and might only temporarily need or even fully avoid a physical manifestation. They will manage funds, pay humans for labour, possess objects and, if needed, create additional entities. So far, law enforcement has ignored these digital entities’ existence and tried to get to the natural persons “behind”, in the case of DAOs, the token holders, the miners or the developers. This was the case with the SEC’s DAO report considering the curators as the securities promoters as well as with some legal scholar’s assertion that the DAO itself was to be treated as a general partnership with every natural or legal person participating being jointly and severally liable for the incurred damages.18

However, simply looking through these entities to find responsible natural persons by “piercing the digital veil”, so to say, will prove to be increasingly insufficient with the ever-growing level of autonomy these digital entities reach. In order for the legal system to properly function, we need accountability of autonomous agents. It is therefore only consequential to demand the introduction of digital personhood.

We should reject carbon-chauvinism and the common view that our intelligent machines will always be our unconscious slaves.” - Max Tegmark 19

About

Ozan Polat is Co-Founder of Dezentrum, a Zurich-based think tank focusing on decentralization and its implications for society. He is also Co-Founder of Trust Square AG, a blockchain hub in the heart of Zurich.

Benedikt Schuppli is Co-Founder of Lexon, a programming language and ecosystem for legal smart contracts. He’s admitted to the bar in Switzerland. He is a contributor at Dezentrum and also a researcher at the Blockchain Research Institute. He has published on smart contracts and ICO regulation.

References

[1] One may even ask whether it actually lies within our competence to make such a conscious expansion. In the end, the normative power of the de facto may inevitably do so as digital persons become ubiquitous in our daily lives.

[2] Karl Friedrich von Savigny, System des heutigen Römischen Rechts, 1840, p. 236.

[3] Citizens United v. Federal Election Commission, 558 U.S. 310 (2010); BGE 95 II 481, E.4.

[4] Leviticus 25:23.

[5] Deuteronomy 10:12.

[6] See Yuval Harari’s Homo Deus for a more in-depth exploration of this notion.

[7] Max Tegmark, Life 3.0, p.109

[8] https://en.oxforddictionaries.com/definition/autonomy

[9] Defining the process of “forming one’s own will” is definitely an important step in assessing autonomy. However, this discussion would go beyond the scope of this text and shall be done in a following post.

[10] http://okhaos.com/plantoids/

[11] A distinction made as early as 2011 in the bitcointalk.org forums by user gmaxwell and picked up again by Berlin-based research group Terra0.

[12] https://www.edge.org/response-detail/27126

[13] http://plantoid.org

[14] We use the word “currently” to leave room for future body-independent manifestations of the human mind as proposed by e.g. Ray Kurzweil in regard to “brain uploads”.

[15] https://wwwwwwwwwwwwwwwwwwwwww.bitnik.org/r/2015-04-15-random-darknet-shopper-free/

[16] https://terra0.org

[17] As quoted in https://terra0.org/assets/pdf/terra0_white_paper_2016.pdf

[18] Stephen D. Palley, How to Sue a Decentralized Autonomous Organization, COINDESK (Mar. 20, 2016), http://www.coindesk.com/how-to-sue-a-decentralized-autonomous-organization/, see also Carla R. Reyes, “If Rockefeller were a Coder”, p. 1.

[19] https://www.edge.org/response-detail/27126
Artikel teilen