Examples
Worked examples
- Is an instance
An author who used a chatbot to improve the phrasing of a discussion section describes this use, including the tool and version, in the manuscript's Acknowledgments section rather than naming the tool as a co-author, consistent with ICMJE's routing of writing-assistance disclosures.
- Is an instance
An author who used an AI tool to help generate a figure from a dataset reports that use in the Methods section, since ICMJE directs AI use in data collection, analysis, or figure generation to Methods rather than Acknowledgments.
Counter-examples
Looks similar, but isn't
- Not an instance
A manuscript byline listing 'ChatGPT' or 'GPT-4' as a co-author is non-compliant with ICMJE's Recommendations, since a chatbot cannot be accountable for the work or approve the final submitted version, which ICMJE treats as a precondition for authorship regardless of the tool's contribution.
Editorial commentary
Proof-of-concept funding sits at a specific point in the research-to-commercialization pipeline: after a discovery has been made — and often after an invention disclosure has already been filed with a technology-transfer office (TTO) — but before that discovery is far enough along to support a licence, a spinout, or outside investment on its own. Its purpose is narrow and consistent across the many schemes that use the label: reduce technical and/or commercial risk on a specific, already-identified result, using milestone-based, typically non-dilutive funding, so that the technology is more attractive to a licensee, investor, or follow-on funder.
How proof-of-concept funding differs from seed funding
The two categories are often confused because both provide small, early awards, but they answer different questions and sit at different points in a research idea’s life:
- Seed funding asks: does this idea deserve initial support to generate the preliminary data or pilot results needed to compete for a larger research grant? It is typically awarded by institutions, learned societies, or research funders and is judged against research-progress criteria.
- Proof-of-concept funding asks: is this already-existing result — a finding, an invention, a prototype — technically and/or commercially viable enough to license, spin out, or attract outside investment? It typically requires a specific underlying discovery (often a filed invention disclosure) as the starting point, and progress is judged against technology-readiness or commercialization milestones rather than research-progress criteria.
A useful shorthand: seed funding helps a research idea become fundable; proof-of-concept funding helps an already-funded discovery become licensable, investable, or spinout-ready. See CASRAI’s seed funding entry for the earlier-stage mechanism, and the technology readiness level (TRL) entry for the milestone framework proof-of-concept programs frequently report against.
Where proof-of-concept funding comes from
Because the underlying need — de-risking a specific discovery before it can be licensed, spun out, or externally invested in — is common across research systems, proof-of-concept funding is offered under many different names and administrative structures rather than a single standardized program:
- Funder-run schemes tied to a prior award. The ERC Proof of Concept Grant is a well-documented example: a EUR 150,000, up-to-18-month top-up award, open only to principal investigators who currently hold or have held one of the European Research Council’s four main frontier-research grants, used specifically to explore the commercial or societal application potential of results already generated under that funded project.
- University-run internal funds, often called PoC funds, gap funds, or commercialization funds. Administered by or alongside the institution’s TTO, these typically require an existing invention disclosure, fund a small number of defined technical and business-development milestones (often in the low tens of thousands of dollars), and run on a fixed application-to-decision cycle of roughly a year or less.
- National/agency translational programs. In the US, NIH’s REACH (Research Evaluation and Commercialization Hubs) program funds proof-of-concept and early commercialization support — prototyping, market validation, mentoring — for results emerging from NIH-funded biomedical research; NSF’s I-Corps curriculum performs a related but distinct function, funding structured customer-discovery activity (a minimum number of stakeholder interviews) rather than technical prototype development itself. Most university PoC/gap funds and I-Corps are complementary rather than duplicative: many institutional PoC funds require or strongly recommend I-Corps completion first, since the customer-discovery evidence I-Corps produces often informs which technologies a PoC fund should prioritize.
See CASRAI’s guide on university innovation and accelerator programs for how these instruments — PoC/gap funds, I-Corps, REACH, and venture accelerators — typically fit together inside a single institution’s commercialization pipeline, and the guide on funding options for a university spinout for how proof-of-concept-stage outcomes typically feed into later spinout financing.
What proof-of-concept funding is typically used for
- Building or refining a working prototype to demonstrate technical feasibility beyond a laboratory-scale result
- Freedom-to-operate screening or provisional patent filing, often run in parallel with or immediately after the funded activity
- Market or stakeholder validation — structured interviews with potential licensees, industry partners, or investors
- Generating the specific data package a licensing negotiation or investor due-diligence process will ask for
Why the distinction matters for research administrators
Confusing the two categories has practical consequences. A researcher pointed toward a proof-of-concept fund when what they actually need is preliminary pilot data for a future grant application will find the eligibility criteria a poor fit — most PoC schemes explicitly require an existing invention or result as the starting point and will not fund open-ended exploratory research. Conversely, a TTO evaluating whether a technology is ready for a PoC award, rather than still needing basic seed-stage development, typically looks for signals such as a filed invention disclosure, a defined technical result, and a plausible commercialization or licensing pathway — the same signals that distinguish PoC funding from seed funding in the first place. Tracking which category a given internal or external funding opportunity actually falls into also matters for how an institution routes the application (research office versus TTO) and how the resulting award is classified in a research information system.
Related terms
- Seed Funding
- ERC Proof of Concept Grant
- Technology Readiness Level (TRL)
- University Innovation and Accelerator Programs
- Funding Options for a University Spinout
Frequently Asked Questions
Does ICMJE allow AI tools like ChatGPT to be listed as an author?
No. ICMJE states that AI tools cannot be authors because they cannot be held responsible for the accuracy, integrity, and originality of the work, and cannot give the approvals that ICMJE’s own authorship criteria require of every author.
When did ICMJE introduce its generative AI policy?
ICMJE added its guidance on generative AI at Section II.A.4 of its Recommendations for the Conduct, Reporting, Editing, and Publication of Scholarly Work in Medical Journals, in the May 2023 update.
Where should authors disclose generative AI use in a manuscript?
It depends on how the tool was used: AI assistance with writing, editing, or proofreading is described in the Acknowledgments section, while AI used to collect or analyze data, or to help generate figures, is reported in the Methods section instead.
Are authors still accountable for content produced with AI assistance?
Yes. ICMJE holds human authors fully responsible for any submitted material involving AI-assisted technologies and expects them to carefully review and edit AI-generated output, since such tools can produce authoritative-sounding text that is incorrect, incomplete, or biased.
Is ICMJE’s generative AI policy legally binding on journals?
No. ICMJE’s Recommendations are guidance that a journal can choose to follow or adapt, not a binding regulation, which is why individual publishers still maintain their own AI policies that reference or extend it.
How does ICMJE’s policy differ from a publisher’s own generative AI policy, such as Nature’s or IEEE’s?
ICMJE’s Recommendations are a widely cited baseline followed or referenced by thousands of biomedical and health-science journals, whereas policies such as the Nature Portfolio AI Policy or the IEEE Generative AI Policy build on similar underlying norms but are written and enforced independently by those publishers.
Machine-readable encodings
Use in your systems
<role vocab="credit"
vocab-identifier="https://casrai.org/dictionary/"
vocab-term="ICMJE Generative AI Policy"
vocab-term-identifier="https://casrai.org/dictionary/term/icmje-generative-ai-policy" />{
"@context": "https://schema.org",
"@type": "DefinedTerm",
"@id": "https://casrai.org/dictionary/term/icmje-generative-ai-policy",
"name": "ICMJE Generative AI Policy",
"identifier": "https://casrai.org/dictionary/term/icmje-generative-ai-policy",
"description": "The International Committee of Medical Journal Editors (ICMJE) sets out its rules on generative AI in Section II.A.4 of its Recommendations for the Conduct, Reporting, Editing, and Publication of Scholarly Work in Medical Journals, added in the May 2023 update. The policy rests on three linked rules: (1) chatbots and other AI tools cannot be listed as authors because they cannot take responsibility for the accuracy, integrity, and originality of the work and cannot approve a final version for submission, so they fail ICMJE's own authorship criteria by definition; (2) at submission, authors must disclose whether AI-assisted technologies were used, with the disclosure location depending on how the tool was used — writing, editing, or proofreading assistance is described in the Acknowledgments section, while use of AI to collect or analyze data or to generate figures is reported in the Methods section; and (3) human authors remain fully responsible for any submitted material produced with AI assistance, including reviewing and editing AI output carefully, because such tools can generate authoritative-sounding text that is incorrect, incomplete, or biased.",
"inDefinedTermSet": "https://casrai.org/dictionary/domain/genai-disclosure#set",
"url": "https://casrai.org/dictionary/term/icmje-generative-ai-policy",
"sameAs": [],
"license": "https://creativecommons.org/licenses/by/4.0/",
"publisher": {
"@id": "https://casrai.org/#organization"
},
"dateModified": "2026-08-15T04:38:06",
"inLanguage": "en"
}






