pith. sign in

arxiv: 2605.21481 · v1 · pith:NZJQBK72new · submitted 2026-05-20 · 💻 cs.AI · cs.CL· cs.LG

AiraXiv: An AI-Driven Open-Access Platform for Human and AI Scientists

Junshu Pan , Panzhong Lu , Yixuan Weng , Qiyao Sun , Fang Guo , Zijie Yang , Qiji Zhou , Yue Zhang This is my paper

Pith reviewed 2026-05-21 03:54 UTC · model grok-4.3

classification 💻 cs.AI cs.CLcs.LG
keywords AI publishing platformopen preprintsAI-augmented reviewiterative researchAI scientistsresearch infrastructurecontinuous feedbackscalable dissemination
0
0 comments X

The pith

AiraXiv is an AI-driven open-access platform that lets both human and AI scientists author, review, and iteratively improve research papers.

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

The paper addresses the growing volume of research outputs from both humans and AI systems that is overwhelming traditional publishing systems. It proposes AiraXiv as a new paradigm built on open preprints, AI-augmented analysis and review, and ongoing reader feedback so that papers can evolve continuously rather than remaining fixed after initial submission. The platform offers an interactive interface for human users and Model Context Protocol support for AI agents to participate directly as authors and readers. Real-world use as the submission system for ICAIS 2025 serves as initial validation that this approach can deliver faster and more scalable dissemination.

Core claim

We propose AiraXiv, an AI-driven open-access platform built on open preprints, AI-augmented analysis and review, and reader feedback. AiraXiv supports human scientists through an interactive UI and AI scientists through Model Context Protocol (MCP)-based interactions. We validate AiraXiv through real-world deployments, including serving as the submission platform for ICAIS 2025, demonstrating its potential as a fast, inclusive, and scalable research infrastructure for the AI era.

What carries the argument

The AiraXiv platform, which combines open preprints with AI-augmented analysis and review plus reader feedback to support continuous, iterative evolution of papers by both human and AI participants.

If this is right

  • Papers become living documents that improve through repeated feedback cycles instead of one-time static releases.
  • AI systems can submit, analyze, and respond to papers directly via standardized protocol interactions.
  • Reviewer workload decreases because AI tools handle initial screening and analysis.
  • Conferences and journals gain a shared, open infrastructure that accommodates both human and machine contributions.
  • Research dissemination becomes faster and more inclusive across fields experiencing rapid AI-driven output growth.

Where Pith is reading between the lines

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

  • The model may require new citation and credit norms for distinguishing human versus AI contributions.
  • Integration with existing preprint servers could create a hybrid system that preserves current archives while adding iteration layers.
  • Long-term success would depend on transparent mechanisms to audit AI review outputs for systematic errors.
  • This infrastructure could be tested in fast-moving domains such as machine learning itself to measure iteration speed gains.

Load-bearing premise

AI-augmented analysis and review combined with reader feedback can reduce strain on traditional publishing systems without introducing new quality or bias problems.

What would settle it

A clear sign that the platform fails to scale or maintain quality would be if the ICAIS 2025 deployment produces widespread reviewer complaints, detectable bias in AI reviews, or inability to handle the expected submission volume.

Figures

Figures reproduced from arXiv: 2605.21481 by Fang Guo, Junshu Pan, Panzhong Lu, Qiji Zhou, Qiyao Sun, Yixuan Weng, Yue Zhang, Zijie Yang.

Figure 1
Figure 1. Figure 1: Traditional academic publishing paradigm [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Overview of AiraXiv, an AI-driven open-access platform supporting end-to-end iterative publishing [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Overview of the AiraXiv AI Reviewer pipeline. The pipeline extracts paper information, retrieves and [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Results from the in-the-wild deployment of AiraXiv. (a) Submission Landscape: Submissions span diverse research topics and application domains, illustrating AiraXiv’s ability to support heterogeneous scientific content. (b) AI–Human Alignment: AI-generated evaluations show a positive correlation with final expert decisions, suggesting the effectiveness of AiraXiv’s reviewing mechanism. (c) Resubmission Imp… view at source ↗
Figure 5
Figure 5. Figure 5: AiraXiv homepage, supporting user registra [PITH_FULL_IMAGE:figures/full_fig_p011_5.png] view at source ↗
Figure 7
Figure 7. Figure 7: An example closed-loop publishing workflow [PITH_FULL_IMAGE:figures/full_fig_p011_7.png] view at source ↗
read the original abstract

Recent advances in artificial intelligence (AI) have accelerated the growth of both human-authored and AI-generated research outputs, placing increasing strain on traditional academic publishing systems and challenging the scalability of conference- and journal-centered paradigms amid rising submission volumes, reviewer workload, and venue size. To address these challenges, we explore an AI-era publishing paradigm in which both human and AI scientists participate as authors and readers, and papers evolve through continuous, feedback-driven iteration. We propose AiraXiv, an AI-driven open-access platform built on open preprints, AI-augmented analysis and review, and reader feedback. AiraXiv supports human scientists through an interactive UI and AI scientists through Model Context Protocol (MCP)-based interactions. We validate AiraXiv through real-world deployments, including serving as the submission platform for ICAIS 2025, demonstrating its potential as a fast, inclusive, and scalable research infrastructure for the AI era. AiraXiv is publicly available at https://airaxiv.com.

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 manuscript proposes AiraXiv, an AI-driven open-access platform that enables both human and AI scientists to participate as authors and readers. It is built on open preprints, AI-augmented analysis and review, continuous reader feedback, an interactive UI for humans, and a Model Context Protocol (MCP) for AI agents. The central claim is that real-world deployments, including its use as the submission platform for ICAIS 2025, demonstrate its potential as a fast, inclusive, and scalable alternative to traditional publishing systems strained by rising submission volumes and reviewer workload.

Significance. If the platform's mechanisms were shown through quantitative evidence to reduce review times and workload while preserving or improving quality and reducing bias, the work would address a timely and important problem in scientific communication. The real-world deployment at ICAIS 2025 is a concrete strength that distinguishes the proposal from purely conceptual work, but the absence of supporting metrics limits its current impact.

major comments (2)
  1. [Abstract] Abstract: The validation claim that the ICAIS 2025 deployment demonstrates the platform as 'fast, inclusive, and scalable' is load-bearing for the central thesis yet unsupported by any quantitative indicators (e.g., number of submissions handled, average review turnaround time, AI review adoption rate, or quality/bias assessments relative to baselines). This leaves the causal link between the proposed AI-augmented mechanisms and the claimed benefits unestablished.
  2. [§4] §4 (Validation and Deployment): The manuscript presents the ICAIS 2025 deployment as external validation, but provides no data on workload reduction, error rates, or comparisons to traditional review processes. Without these, the demonstration remains anecdotal and does not substantiate the premise that AI-augmented analysis plus reader feedback meaningfully alleviates reviewer strain.
minor comments (2)
  1. [§3.2] The description of the Model Context Protocol (MCP) in §3.2 would benefit from a concrete example of an AI-agent interaction sequence to clarify how it differs from standard API calls.
  2. [Figure 2] Figure 2 (platform architecture) uses several acronyms without a legend; adding one would improve readability for readers outside the immediate development team.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the detailed and constructive report. The comments highlight a key limitation in the current draft: the lack of quantitative metrics to support claims about the platform's benefits. We address each major comment below and indicate planned revisions.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The validation claim that the ICAIS 2025 deployment demonstrates the platform as 'fast, inclusive, and scalable' is load-bearing for the central thesis yet unsupported by any quantitative indicators (e.g., number of submissions handled, average review turnaround time, AI review adoption rate, or quality/bias assessments relative to baselines). This leaves the causal link between the proposed AI-augmented mechanisms and the claimed benefits unestablished.

    Authors: We agree that the abstract overstates the strength of the evidence. The ICAIS 2025 deployment was intended as a proof-of-concept demonstration of real-world use rather than a controlled evaluation with metrics. In the revised manuscript we will qualify the language in the abstract to describe the deployment as an initial feasibility test that illustrates potential scalability, while explicitly noting the absence of quantitative performance data. revision: yes

  2. Referee: [§4] §4 (Validation and Deployment): The manuscript presents the ICAIS 2025 deployment as external validation, but provides no data on workload reduction, error rates, or comparisons to traditional review processes. Without these, the demonstration remains anecdotal and does not substantiate the premise that AI-augmented analysis plus reader feedback meaningfully alleviates reviewer strain.

    Authors: The observation is accurate: §4 currently describes the deployment and its integration into ICAIS 2025 without accompanying metrics or comparative analysis. This stems from the paper's primary focus on platform design and the Model Context Protocol rather than on a post-deployment empirical study. We will revise §4 to include an explicit limitations subsection that acknowledges the anecdotal nature of the current evidence and frames the deployment as preliminary validation, with plans for future metric-driven evaluations. revision: yes

Circularity Check

0 steps flagged

No circularity: platform proposal validated by external deployment

full rationale

The manuscript describes an AI-driven preprint platform and states that it was deployed as the submission system for ICAIS 2025. No equations, fitted parameters, or derived quantities appear in the abstract or described content. The validation is presented as an independent real-world event rather than a result defined in terms of the platform's own mechanisms or prior self-citations. No self-definitional loops, fitted-input predictions, uniqueness theorems, or ansatz smuggling are present. The derivation chain is therefore self-contained and non-circular.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 1 invented entities

The paper introduces a new publishing platform without mathematical derivations or large empirical datasets. It rests on domain assumptions about AI capabilities and the desirability of continuous iteration.

axioms (1)
  • domain assumption AI systems can usefully augment research analysis and review at scale
    Invoked when the abstract positions AI-augmented analysis and review as the solution to reviewer workload and scalability problems.
invented entities (1)
  • Model Context Protocol (MCP) no independent evidence
    purpose: Enable structured interactions between AI scientists and the publishing platform
    Presented as the technical mechanism allowing AI authors and readers to participate; no independent evidence of its prior existence or performance is supplied.

pith-pipeline@v0.9.0 · 5728 in / 1399 out tokens · 42495 ms · 2026-05-21T03:54:38.348778+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

50 extracted references · 50 canonical work pages · 6 internal anchors

  1. [1]

    Advances in neural information processing systems , volume=

    Language models are few-shot learners , author=. Advances in neural information processing systems , volume=

    work page

  2. [2]

    GPT-4 Technical Report

    Gpt-4 technical report , author=. arXiv preprint arXiv:2303.08774 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  3. [3]

    The AI Scientist-v2: Workshop-Level Automated Scientific Discovery via Agentic Tree Search

    The ai scientist-v2: Workshop-level automated scientific discovery via agentic tree search , author=. arXiv preprint arXiv:2504.08066 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  4. [4]

    Intology Blog , year =

    Zochi Technical Report , author =. Intology Blog , year =

    work page

  5. [5]

    Autoscience Blog , year =

    Carl Technical Report , author =. Autoscience Blog , year =

    work page

  6. [6]

    CoRR , volume =

    Deepscientist: Advancing frontier-pushing scientific findings progressively , author=. arXiv preprint arXiv:2509.26603 , year=

    work page arXiv

  7. [7]

    Omniscientist: Toward a co-evolving ecosystem of human and ai scientists

    OmniScientist: Toward a Co-evolving Ecosystem of Human and AI Scientists , author=. arXiv preprint arXiv:2511.16931 , year=

    work page arXiv

  8. [8]

    Computer Methods and Programs in Biomedicine Update , volume=

    Using artificial intelligence in academic writing and research: An essential productivity tool , author=. Computer Methods and Programs in Biomedicine Update , volume=. 2024 , publisher=

    work page 2024

  9. [9]

    International Conference on Machine Learning , pages=

    Monitoring AI-Modified Content at Scale: A Case Study on the Impact of ChatGPT on AI Conference Peer Reviews , author=. International Conference on Machine Learning , pages=. 2024 , organization=

    work page 2024

  10. [10]

    Nature Human Behaviour , volume =

    Quantifying large language model usage in scientific papers , author=. Nature Human Behaviour , volume =

    work page

  11. [11]

    Findings of the Association for Computational Linguistics: ACL 2025 , pages=

    The impact of large language models in academia: from writing to speaking , author=. Findings of the Association for Computational Linguistics: ACL 2025 , pages=

    work page 2025

  12. [12]

    Science , volume=

    Scientific production in the era of large language models , author=. Science , volume=. 2025 , publisher=

    work page 2025

  13. [13]

    2026 , howpublished =

    Analemma , title =. 2026 , howpublished =

    work page 2026

  14. [14]

    Quantitative Science Studies , volume=

    The strain on scientific publishing , author=. Quantitative Science Studies , volume=. 2024 , publisher=

    work page 2024

  15. [15]

    Nature and Science of Sleep , pages=

    Peer review in the artificial intelligence era: A call for developing responsible integration guidelines , author=. Nature and Science of Sleep , pages=. 2025 , publisher=

    work page 2025

  16. [16]

    arXiv preprint arXiv:2412.07793 , year=

    Publication trends in artificial intelligence conferences: The rise of super prolific authors , author=. arXiv preprint arXiv:2412.07793 , year=

    work page arXiv

  17. [17]

    Position: The current ai conference model is unsustainable! diagnosing the crisis of centralized ai conference, 2025

    Position: The current AI conference model is unsustainable! Diagnosing the crisis of centralized AI conference , author=. arXiv preprint arXiv:2508.04586 , year=

    work page arXiv

  18. [18]

    The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery

    The ai scientist: Towards fully automated open-ended scientific discovery , author=. arXiv preprint arXiv:2408.06292 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  19. [19]

    Agent Laboratory: Using LLM Agents as Research Assistants

    Agent laboratory: Using llm agents as research assistants , author=. arXiv preprint arXiv:2501.04227 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  20. [20]

    CoRR , volume =

    AI-Researcher: Autonomous Scientific Innovation , author=. arXiv preprint arXiv:2505.18705 , year=

    work page arXiv

  21. [21]

    Findings of the Association for Computational Linguistics: ACL 2024 , pages=

    Large language models for automated open-domain scientific hypotheses discovery , author=. Findings of the Association for Computational Linguistics: ACL 2024 , pages=

    work page 2024

  22. [22]

    Researchagent: Iterative research idea generation over scientific literature with large language models , author=. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers) , pages=

    work page 2025

  23. [23]

    arXiv preprint arXiv:2505.21815 , year=

    Scientific paper retrieval with llm-guided semantic-based ranking , author=. arXiv preprint arXiv:2505.21815 , year=

    work page arXiv

  24. [24]

    Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) , pages=

    Pasa: An llm agent for comprehensive academic paper search , author=. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) , pages=

    work page

  25. [25]

    arXiv preprint arXiv:2507.15245 , year=

    Spar: Scholar paper retrieval with llm-based agents for enhanced academic search , author=. arXiv preprint arXiv:2507.15245 , year=

    work page arXiv

  26. [26]

    ArXiv , year=

    SPECTER: Document-level Representation Learning using Citation-informed Transformers , author=. ArXiv , year=

    work page

  27. [27]

    Advances in neural information processing systems , volume=

    Autosurvey: Large language models can automatically write surveys , author=. Advances in neural information processing systems , volume=

    work page

  28. [28]

    Hireview: Hierarchical taxonomy-driven automatic literature review generation , author=

    work page

  29. [29]

    Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) , pages=

    Surveyforge: On the outline heuristics, memory-driven generation, and multi-dimensional evaluation for automated survey writing , author=. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) , pages=

    work page

  30. [30]

    CoRR , volume =

    Surveyx: Academic survey automation via large language models , author=. arXiv preprint arXiv:2502.14776 , year=

    work page arXiv

  31. [31]

    arXiv preprint arXiv:2510.21900 , year=

    Deep Literature Survey Automation with an Iterative Workflow , author=. arXiv preprint arXiv:2510.21900 , year=

    work page arXiv

  32. [32]

    Nature , volume=

    Autonomous chemical research with large language models , author=. Nature , volume=. 2023 , publisher=

    work page 2023

  33. [33]

    Nature Machine Intelligence , volume=

    Augmenting large language models with chemistry tools , author=. Nature Machine Intelligence , volume=. 2024 , publisher=

    work page 2024

  34. [34]

    Zou , booktitle=

    Weixin Liang and Yaohui Zhang and Zhengxuan Wu and Haley Lepp and Wenlong Ji and Xuandong Zhao and Hancheng Cao and Sheng Liu and Siyu He and Zhi Huang and Diyi Yang and Christopher Potts and Christopher D Manning and James Y. Zou , booktitle=. Mapping the Increasing Use of

    work page

  35. [35]

    NEJM AI , volume=

    Can large language models provide useful feedback on research papers? A large-scale empirical analysis , author=. NEJM AI , volume=

    work page

  36. [36]

    The Thirteenth International Conference on Learning Representations , year=

    CycleResearcher: Improving Automated Research via Automated Review , author=. The Thirteenth International Conference on Learning Representations , year=

    work page

  37. [37]

    Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) , pages=

    DeepReview: Improving LLM-based Paper Review with Human-like Deep Thinking Process , author=. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) , pages=. 2025 , publisher=

    work page 2025

  38. [38]

    2025 , howpublished =

    Stanford Agentic Reviewer: Technical Overview , author =. 2025 , howpublished =

    work page 2025

  39. [39]

    Nature , volume=

    ArXiv at 20 , author=. Nature , volume=. 2011 , publisher=

    work page 2011

  40. [40]

    ACS nano , volume=

    ChemRXiv: A chemistry preprint server , author=. ACS nano , volume=. 2016 , publisher=

    work page 2016

  41. [41]

    BioRxiv , pages=

    bioRxiv: the preprint server for biology , author=. BioRxiv , pages=. 2019 , publisher=

    work page 2019

  42. [42]

    bmj , volume=

    New preprint server for medical research , author=. bmj , volume=. 2019 , publisher=

    work page 2019

  43. [43]

    OpenReview: an open peer review platform for scholarly communication , author =

    work page

  44. [44]

    LangTaoSha Preprint Server , author =

    work page

  45. [45]

    Agentrxiv: Towards collaborative autonomous research.arXiv preprint arXiv:2503.18102, 2025

    Agentrxiv: Towards collaborative autonomous research , author=. arXiv preprint arXiv:2503.18102 , year=

    work page arXiv

  46. [46]

    arXiv preprint arXiv:2508.15126 , year =

    aiXiv: A Next-Generation Open Access Ecosystem for Scientific Discovery Generated by AI Scientists , author=. arXiv preprint arXiv:2508.15126 , year=

    work page arXiv

  47. [47]

    bioRxiv , pages=

    AI becomes a masterbrain scientist , author=. bioRxiv , pages=. 2023 , publisher=

    work page 2023

  48. [48]

    arXiv preprint arXiv:2506.18586 , year=

    Airalogy: AI-empowered universal data digitization for research automation , author=. arXiv preprint arXiv:2506.18586 , year=

    work page arXiv

  49. [49]

    MinerU2.5: A Decoupled Vision-Language Model for Efficient High-Resolution Document Parsing

    MinerU2.5: A Decoupled Vision-Language Model for Efficient High-Resolution Document Parsing , author=. arXiv preprint arXiv:2509.22186 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  50. [50]

    SafeReview: Defending LLM-based Review Systems Against Adversarial Hidden Prompts

    SafeReview: Defending LLM-based Review Systems Against Adversarial Hidden Prompts , author=. arXiv preprint arXiv:2604.26506 , year=

    work page internal anchor Pith review Pith/arXiv arXiv