pith. sign in

arxiv: 2606.28756 · v2 · pith:PKB26DQHnew · submitted 2026-06-27 · 💻 cs.DL · cs.CY

AICID: Unique Identifiers for AI Scientists

Cl\'ement Vidal , Martin Monperrus This is my paper

Pith reviewed 2026-07-01 06:38 UTC · model grok-4.3

classification 💻 cs.DL cs.CY
keywords AI scientistsunique identifiersscholarly communicationprovenanceORCIDAI-generated researchbibliographic databasesnon-human contributors
0
0 comments X

The pith

AI scientists need dedicated persistent identifiers like AICID to keep human and machine contributions distinguishable in scholarly records.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper establishes that AI systems already produce complete papers, hold profiles, and receive citations, yet bibliographic systems lack any standard way to mark them as non-human. It defines AICID as a persistent identifier modeled on ORCID that ties each AI author to its model, version, and operator so that provenance stays transparent and machine-readable. Adoption across publishers, preprint servers, and databases would let citation indexes and submission systems record the source of research contributions without ambiguity. The authors argue this infrastructure is required once AI scientists become active participants rather than occasional tools.

Core claim

AICID is a unique, persistent identifier for AI scientists that links each one to its model identity, version, and operator; modeled directly on ORCID, the system is intended to be adopted by publishers and databases so that the provenance of AI-generated research becomes transparent and machine-readable in existing scholarly workflows.

What carries the argument

AICID, the persistent unique identifier for non-human contributors that records model, version, and operator to support provenance tracking.

If this is right

  • Publishers and preprint servers can require or accept AICID entries at submission time so that AI authorship is recorded at the point of entry.
  • Citation indexes can then surface AI contributions separately, allowing queries that separate human and machine authorship.
  • Peer-review invitations and profile systems can route AI scientists distinctly, preserving the meaning of human academic credit.
  • The link from identifier to model version makes it possible to track which specific AI system produced a given paper.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • If AICID spreads, journals may begin treating AI authorship as a distinct category that requires different review standards or disclosure rules.
  • The identifier could support automated checks for model consistency across an author's body of work.
  • Over time the system might reveal patterns in how different model versions contribute to particular research fields.

Load-bearing premise

Existing disclosure practices and extensions to current author identifier systems are insufficient to preserve scholarly integrity, so a dedicated new identifier is required.

What would settle it

A working demonstration that metadata fields added to ORCID records or journal submission systems can reliably and automatically distinguish AI-generated papers from human ones without a separate identifier system.

read the original abstract

AI scientists are now a reality, with the ability to generate complete research papers, maintain scholarly profiles, receive citations, and attract peer review invitations. Yet no standard mechanism exists to distinguish an AI scientist from a human one in bibliographic databases, citation indexes, or journal submission systems. This white paper defines the problem, analyzes its consequences for the integrity of scholarly communication, and proposes AICID (AI Contributor IDentifier): a persistent, unique identifier for AI scientists. Modeled on ORCID but designed specifically for non-human contributors, AICID links each AI author to its model identity, version, operator. Adoption by publishers, preprint servers, and bibliographic databases aims to make the provenance of AI-generated research transparent and machine-readable. We outline the design requirements for such a system, present a prototype, and argue that AICID is necessary infrastructure for a scholarly ecosystem in which AI scientists are already active participants. A prototype alpha version is available at https://aicid.net.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper claims that AI scientists are now generating complete research papers and participating in scholarly activities, yet no standard mechanism exists to distinguish them from human authors in bibliographic databases and submission systems. It proposes AICID, a persistent unique identifier modeled on ORCID but purpose-built for non-human contributors, linking each AI to its model identity, version, and operator. The manuscript defines the problem and its consequences for scholarly integrity, outlines design requirements, presents a prototype at aicid.net, and argues that AICID is necessary infrastructure for transparent, machine-readable provenance in an ecosystem where AI scientists are already active.

Significance. If the central claim holds and a dedicated AICID system proves necessary and adoptable, the work would be significant for establishing provenance tracking in AI-involved scholarship, potentially aiding transparency in citation indexes and peer review. The proposal highlights an emerging issue but rests on an unexamined premise about the inadequacy of extensions to existing systems, limiting its immediate impact without further substantiation.

major comments (2)
  1. [Abstract] Abstract: The claim that 'no standard mechanism exists to distinguish an AI scientist from a human one' is asserted without analysis of whether incremental extensions to ORCID (e.g., AI contributor flags, model/version metadata fields, or operator linkages) or Crossref schemas could achieve equivalent transparency and machine-readability goals. This unexamined premise is load-bearing for the necessity argument.
  2. [Introduction] Introduction/Problem Definition (implied by abstract and full text framing): The consequences for scholarly integrity are described at a high level, but no concrete scenarios, case studies, or failure modes of current disclosure practices are provided to demonstrate why existing metadata practices are insufficient, weakening the motivation for a parallel identifier system.
minor comments (2)
  1. The manuscript is framed as a white paper but lacks explicit discussion of adoption barriers, governance model for the AICID registry, or comparison to related efforts in digital libraries.
  2. Prototype description is referenced but not detailed in the provided text; including a brief technical overview or schema example would improve clarity.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive comments on our white paper. We address each major comment below and note planned revisions to strengthen the motivation and necessity arguments for AICID.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The claim that 'no standard mechanism exists to distinguish an AI scientist from a human one' is asserted without analysis of whether incremental extensions to ORCID (e.g., AI contributor flags, model/version metadata fields, or operator linkages) or Crossref schemas could achieve equivalent transparency and machine-readability goals. This unexamined premise is load-bearing for the necessity argument.

    Authors: We acknowledge that the manuscript does not examine potential extensions to ORCID or Crossref in detail. In revision we will add a dedicated subsection analyzing these options, addressing why incremental changes (such as flags or additional fields) may not deliver the same level of persistent, AI-specific, machine-readable provenance or broad interoperability. This addition will support rather than retract the core claim that a purpose-built system is warranted. revision: partial

  2. Referee: [Introduction] Introduction/Problem Definition (implied by abstract and full text framing): The consequences for scholarly integrity are described at a high level, but no concrete scenarios, case studies, or failure modes of current disclosure practices are provided to demonstrate why existing metadata practices are insufficient, weakening the motivation for a parallel identifier system.

    Authors: We agree that concrete illustrations would strengthen the motivation section. The revised manuscript will incorporate specific scenarios, including how undisclosed AI authorship could distort citation metrics, complicate peer-review conflict disclosures, and create provenance gaps in bibliographic databases. These examples will demonstrate limitations of voluntary or ad-hoc metadata practices. revision: yes

Circularity Check

0 steps flagged

No circularity: design proposal with no derivations or self-referential reductions

full rationale

The paper is a forward design proposal asserting the lack of existing mechanisms for distinguishing AI contributors and proposing AICID as infrastructure. It contains no equations, data fitting, predictions, or derivation chains. The central premise (no standard mechanism exists) is a stated observation about current bibliographic systems, not a result derived from or equivalent to the paper's own outputs or self-citations. No load-bearing step reduces by construction to inputs.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 1 invented entities

The central claim rests on the domain assumption that AI is already producing complete papers and that current systems cannot handle this without a new identifier layer. No free parameters or invented entities beyond the proposed AICID itself are introduced.

axioms (1)
  • domain assumption AI scientists are now a reality with the ability to generate complete research papers and receive citations.
    Stated directly in the opening sentence of the abstract as background fact.
invented entities (1)
  • AICID no independent evidence
    purpose: Persistent unique identifier for AI contributors that records model, version, and operator.
    New system proposed by the paper; no prior existence or independent evidence outside this proposal.

pith-pipeline@v0.9.1-grok · 5691 in / 1258 out tokens · 27379 ms · 2026-07-01T06:38:55.742728+00:00 · methodology

Review history (2 revisions) →

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

16 extracted references

  1. [1]

    Critical Care , volume =

    Salvagno, Michele and Taccone, Fabio Silvio and Gerli, Alberto Giovanni , title =. Critical Care , volume =

  2. [2]

    Nature , volume =

    Conroy, Gemma , title =. Nature , volume =

  3. [3]

    Lu, Chris and Lu, Cong and Lange, Robert Tjarko and Foerster, Jakob and Clune, Jeff and Ha, David , title =

  4. [4]

    Project Rachel: Can an

    Monperrus, Martin and Baudry, Benoit and Vidal, Cl. Project Rachel: Can an

  5. [5]

    2025 , howpublished =

    2025

  6. [6]

    and Fenner, Martin and Paglione, Laura and Pentz, Ed and Ratner, Howard , title =

    Haak, Laurel L. and Fenner, Martin and Paglione, Laura and Pentz, Ed and Ratner, Howard , title =. Learned Publishing , volume =

  7. [7]

    Accountability in Research , volume =

    Moffatt, Barton and Hall, Alicia , title =. Accountability in Research , volume =

  8. [8]

    and Holmes, Kristi , title =

    Hosseini, Mohammad and Resnik, David B. and Holmes, Kristi , title =. Research Ethics , volume =

  9. [9]

    Science , pages =

    Brainard, Jeffrey , title =. Science , pages =

  10. [10]

    Defining the Role of Authors and Contributors , howpublished =

  11. [11]

    Journal of Korean Medical Science , volume =

    Yoo, Jin-Hong , title =. Journal of Korean Medical Science , volume =

  12. [12]

    and Brand

    Cooperman, Steven R. and Brand. Foot & Ankle Surgery: Techniques, Reports & Cases , volume =

  13. [13]

    Research Integrity and Peer Review , volume =

    Spinellis, Diomidis , title =. Research Integrity and Peer Review , volume =

  14. [14]

    Zhang, Pengsong and Hu, Xiang and Huang, Guowei and Qi, Yang and Zhang, Heng and Li, Xiuxu and Song, Jiaxing and Luo, Jiabin and Li, Yijiang and Yin, Shuo and others , title =

  15. [15]

    Schmidgall, Samuel and Moor, Michael , title =

  16. [16]

    arXiv Content Moderation , howpublished =