Proposal: Paradigm of Legal Autonomous Behaviour (PLAB)

Proposals Round 5
#1

Name of Project: Paradigm of Legal Autonomous Behaviour (PLAB)

Proposal in one sentence: PLAB aims at demonstrating how Decentralised AI systems can be considered autonomous legal subjects capable of determining legal relationships arising by conduct in decentralised-based transactions.

Description of the project and what problem is it solving:

The Problem to solve :

Decentralised AI (DzAI) entails some challenges related to the qualification of the role of DzAI in decentralised-based transactions from the legal viewpoint.

Such challenges descend from two main intertwined issues. The first one leans upon the fact that some decentralised transactions, such as smart contract transactions, may show some technical features that enable the applicability of legal contract regulations. Otherwise, the second issue is linked to the qualification of the legal role that the decentralised AI may potentially assume with respect to such decentralised-based transactions, strictly related to the capacity of DzAI to show a high level of autonomous decision-making ability. Therefore, the main hurdle for legal operators to solve is to identify, interpret and qualify which kind of decentralised interactions may entail legal contract transactions and to what extent the autonomy of DzAI systems to determine and perform these transactions may be crucial in unfolding or affecting these legal relationships.

Project description:

The PLAB is a new legal theory project which endorses how Decentralised AI systems may be recognised as autonomous legal subjects able to determine legal relationships which arise from their conduct expressed by their decision-making process. Starting from the analysis of DzAI technical features, the PLAB can depict the level of autonomy of these systems by means of five (5) areas: learning mechanism, internal state, control, adaptation level and goals. The level and combination of these traits can generate results on how the DzAI is able to produce and affect legal relationships using its own autonomous behaviour. The higher the technical autonomy of the decision-making process is, the more relevant the autonomy from a legal point of view in expressing intentional stances in legal contract transactions. Moreover, the possibility to qualify some smart contracts ‘on-chain’ as legal contracts can, thus, demonstrate how the DzAI may aim to be considered as an autonomous legal subject able to form legal contract transactions by reference of conduct. Therefore, the PLAB results may help legal operators in solving the issue of qualifying the role of DzAI systems in smart contract transactions which are extremely relevant from a legal point of view for contract law regulations.**

Grant Deliverables:

Grant Deliverable 1

Demonstrating how smart contract transactions made ‘on-chain’ may represent a new kind of legal contract made by conduct.

Grant Deliverable 2

By means of the PLAB application, demonstrating how DzAI systems that determine and perform smart contracts ‘on-chain’ may express intentional stances by conduct in the formation of such legal contracts.

Grant Deliverable 3

The application of PLAB to DzAI systems’ features can unfold their technical and legal autonomy in the formation of smart contract transactions and may underpin the proposal of recognising DzAI as an autonomous legal subject.

Squad

Lucia della Ventura is a qualified lawyer in Ireland and Italy. She gained two LLM in IT/IP Law and EU Competition Law, currently a PhD candidate at Trinity College Dublin in Decentralised AI legal regulation. Lucia is also a member of the Technology, Legal and Society (TLS) Research Group at Trinity College Dublin and also an international research fellow at the ISLC Institute of the University of Milan. She works as a lawyer and legal and compliance manager at Haycen, a UK and Irish-based crypto company where she manages the legal side of the Open-Source Sterling Project (OSSP) which aims at building a tokenized British Pound for global commerce.

https://www.linkedin.com/in/luciadellaventuraphd/

1 Like
#2

They can be viewed as legal entities, for example if they are registered as an LLC, but the problem with those legal entities, is that they are not subject to the laws, such that a court may issue an order to restrain them. see e.g. Tornado cash

Therefore if for example some part of the contract is held to be void because it is contrary to law, hypothetically speaking a smart contract for assassination marketplaces that scans death certificates, the “smart” contract cannot be made to comply with an order voiding that contract.

However such a smart contract does not need the principal of “conduct”, which is certainly helpful, but rather the “intent to be bound” by an overt act by engaging with the contract, in addition to the “consideration” which is the transfer of “any thing of value” inherent in the contract.

To bind third parties to the contract, in the instance of a asset swap of some type, the smart contract does uses the legal principals of assignment, and the assignees are liable for the obligations that it took by engaging with the contract, including whatever wrongdoing might have taken place when entering into the contract such as misrepresentation.

In the end, i think that the technology isn’t as relevent, because there are already analogies from case law, if we just substitute the smart contracts / AI systems with LLCs and conspiracy law. However a law journal article that matches the case law by analogy, to the systems that are being deployed, is very much useful for developers who don’t want to get caught up in regulations. However it appears that many governments are trying to carve out new rules for AI, mostly in my opinion because they dislike the reorganization of social relationships (like whether artists are compensated for training data when they are made obsolete), and not because there are any legal loopholes that would allow tortious or criminal conduct to be made non tortious or criminal by virtue of not being a human.