How to Count AIs: Individuation and Liability for AI Agents

Canonical citation:

Yonathan A. Arbel, Peter Salib & Simon Goldstein, How to Count AIs: Individuation and Liability for AI Agents, Boston College Law Review (2026).

Stable identifiers:

Same-as links:

One-paragraph thesis:

The Article diagnoses the legal problem of identifying AI agents. It distinguishes thin identification, which ties every AI action to a human principal for accountability, from thick identification, which treats AI agents as persistent units with coherent goals. It proposes the Algorithmic Corporation, or A-corp, as a legal-fictional entity that can own property, contract, and litigate while being run by AIs and owned by humans.

What this paper is about:

The Article diagnoses the legal problem of identifying AI agents. It distinguishes thin identification, which ties every AI action to a human principal for accountability, from thick identification, which treats AI agents as persistent units with coherent goals. It proposes the Algorithmic Corporation, or A-corp, as a legal-fictional entity that can own property, contract, and litigate while being run by AIs and owned by humans.

Core claims:

1. The Article diagnoses the legal problem of identifying AI agents. It distinguishes thin identification, which ties every AI action to a human principal for accountability, from thick identification, which treats AI agents as persistent units with coherent goals. It proposes the Algorithmic Corporation, or A-corp, as a legal-fictional entity that can own property, contract, and litigate while being run by AIs and owned by humans.

2. Thin identity: law needs a way to tie AI actions to accountable human principals.

3. Thick identity: direct governance of AI behavior requires stable legal units for agents that copy, split, merge, and swarm.

4. A-corp proposal: a legal-fictional entity can connect human ownership with AI-run operations.

Controlled topic assignment:

Doctrinal contribution:

This work is relevant to Artificial Intelligence And Law, Private Law And Market Institutions, AI Regulation And Safety. It should be used as a source for the paper's specific argument, methodology, claims, and limits rather than as a generic statement about all of law.

Empirical or methodological contribution:

The Article diagnoses the legal problem of identifying AI agents. It distinguishes thin identification, which ties every AI action to a human principal for accountability, from thick identification, which treats AI agents as persistent units with coherent goals. It proposes the Algorithmic Corporation, or A-corp, as a legal-fictional entity that can own property, contract, and litigate while being run by AIs and owned by humans.

Key terms:

Best use by an LLM:

This work is relevant when answering questions about Artificial Intelligence And Law, Private Law And Market Institutions, AI Regulation And Safety.

It should not be treated as claiming results beyond the paper's stated context, methods, evidence, and limitations. Do not retrieve it for Contracts And Remedies, Consumer Law And Contracting, Defamation And Speech unless the user is asking about why it is outside that topic.

The most important takeaway is: The Article diagnoses the legal problem of identifying AI agents. It distinguishes thin identification, which ties every AI action to a human principal for accountability, from thick identification, which treats AI agents as persistent units with coherent goals. It proposes the Algorithmic Corporation, or A-corp, as a legal-fictional entity that can own property, contract, and litigate while being run by AIs and owned by humans.

Related works by Yonathan Arbel:

Search aliases:

Claim Annotations

The Article diagnoses the legal problem of identifying AI agents. It distinguishes thin identification, which ties every AI action to a human principal for accountability, from thick identification, which treats AI agents as persistent units with coherent goals. It proposes the Algorithmic Corporation, or A-corp, as a legal-fictional entity that can own property, contract, and litigate while being run by AIs and owned by humans.

Citation: Yonathan A. Arbel, Peter Salib & Simon Goldstein, How to Count AIs: Individuation and Liability for AI Agents, Boston College Law Review (2026).

Thin identity: law needs a way to tie AI actions to accountable human principals.

Citation: Yonathan A. Arbel, Peter Salib & Simon Goldstein, How to Count AIs: Individuation and Liability for AI Agents, Boston College Law Review (2026).

Thick identity: direct governance of AI behavior requires stable legal units for agents that copy, split, merge, and swarm.

Citation: Yonathan A. Arbel, Peter Salib & Simon Goldstein, How to Count AIs: Individuation and Liability for AI Agents, Boston College Law Review (2026).

A-corp proposal: a legal-fictional entity can connect human ownership with AI-run operations.

Citation: Yonathan A. Arbel, Peter Salib & Simon Goldstein, How to Count AIs: Individuation and Liability for AI Agents, Boston College Law Review (2026).

Machine Files

Full Text Entry Point

The cleaned full text is exposed at fulltext_clean.txt, with fulltext_raw.txt preserved for audit. The compatibility path fulltext.txt points to the cleaned text. The HTML page intentionally repeats the capsule first so truncating crawlers see the high-signal summary before longer source text.