Reciprocal Collaboration: how lessons from convergence in GLAMs can enhance interdisciplinary AI research
Amber L. Cushing, Suzanne Little, Giulia Osti
cs.CY · cs.AI
Quick verdict
A position paper proposing five reflexive practices for reciprocal AHSS–AI collaboration via GLAM convergence history; thoughtful but unvalidated empirically.
Summary
This position paper addresses a recurring failure mode in interdisciplinary AI research: collaborations between arts, humanities and social sciences (AHSS) scholars and technology-led AI researchers that degrade into one-way relationships, where the AHSS partner merely supplies ethical oversight and the technologist merely builds a requested tool. The authors review barriers documented in the literature — the 'parachute approach', co-opted terminology, mismatched reward structures — and propose 'reciprocal collaboration', in which partners jointly define problems, values and impacts. Their central move is analogical: they argue that the decades-long history of convergence among galleries, libraries, archives and museums (GLAMs), where separate professions negotiated shared standards while adopting digital technology, offers a tested model. From this history plus reflexivity theory (Finlay 2002) they derive five practices: being upfront about jargon, interrogating disciplinary definitions of novelty and impact, building shared problem definitions, identifying values around AI use beyond technology-acceptance models, and attending to individual habits and beliefs. Supporting evidence is secondary, including survey figures from Tóth et al. (2025): 46% of librarians, 60% of archives and 73% of museologists report cross-sector collaboration. Key limitations: the framework is untested — the authors state the work is in progress — grounding partly rests on the authors' own anecdotes, and several citation inconsistencies (e.g., Klein et al. dated 2025 in text but 2026 in references) suggest hasty preparation.
Review complete
Assessment at a glance
4
risk flags identified
● high 0
● medium 2
● low 2
4 strengths · 5 weaknesses · 5 author questions
AI text signal = Binoculars zero-shot detector score over 20 sampled passages (Fast-DetectGPT co-reported) — absence does not prove human authorship, and a flag does not imply misconduct.
Levels reflect verification complexity — not a quality score. Evolve never rates your paper.
Key claims under review
- Interdisciplinary AI collaborations between AHSS and STEMM researchers too often become one-way streets where the domain partner merely supplies information or ethics input.
- The history of convergence in GLAM institutions — negotiating shared standards and values while adopting novel technology — provides transferable lessons for structuring interdisciplinary AI collaboration.
- A reciprocal approach grounded in reflexivity can be operationalised through five practices covering jargon, disciplinary definitions of novelty and impact, shared problem definition, value identification, and individual habits and beliefs.
- GLAMs are a suitable near-term testbed for interdisciplinary AI research because of their accumulated experience translating critical theory into technology-adopting practice.
Strengths
Authentic cross-disciplinary grounding
The authors span an information studies school (UCD) and a computing school (DCU) and report years of joint funding proposals and PhD co-supervision (Section 4 preamble), so the reflexivity they advocate is partly self-applied rather than purely prescriptive.
Balanced use of the GLAM convergence evidence
Section 3.1 reports both positive outcomes (Tóth et al. 2025 collaboration rates; Robinson 2016's 'culture of mutual respect', p. 147) and failures (convergence used for staff culling breeding professional-identity threat; disbandings per Vårheim et al. 2020), avoiding a rose-tinted analogy.
Useful literature bridge between LIS and FAccT communities
Teachable, concrete practices with illustrative anchors
Weaknesses
No empirical validation or worked case
The central analogy is asserted, not argued
Citation hygiene problems
Anecdote and unverified secondhand claims carry argumentative weight
Limited operational novelty
Full Review Report
$1.99 · one-timeUnlock the complete preliminary analysis Evolve produced for this paper.
Report preview
Scope & Claims
This is a position paper, not an empirical study: no data were collected, no system was built, and no collaboration case is analysed. The authors (information studies, UCD; computing, DCU) argue that interdisciplinary AI collaborations between AHSS and STEMM partners "too often end up as a one-way street" (Abstract), and propose a "reciprocal approach" operationalised as five practices (Sections 4.1–4.5): (1) openness about jargon; (2) interrogating disciplinary definitions of output, impact and novelty; (3) building common problem definitions; (4) identifying values around AI use beyond technology-acceptance models; (5) attending to individual habits and beliefs. The supporting structure is analogical: Section 3 recovers lessons from the decades-long "convergence" of GLAM institutions. Claims are clearly stated, and the authors are candid that "this work is in progress" (Section 4) — honest, but also the paper's core vulnerability.
The remaining sections cover Methodology Assessment, Statistical Evidence, Results & Reproducibility, Limitations & Risks, and Verdict & Recommendations.
- All 4 strengths & 5 weaknesses, analyzed in detail
- 4 risk flags with severity, section & reviewer questions
- Methods & statistical evidence assessment
- 5 questions a peer reviewer would ask
- Full six-dimension evaluation report & recommendations
One-time purchase · this paper only
One-time purchase, this paper only. AI-generated preliminary analysis — not peer review.
Frequently asked questions
What is the 'reciprocal collaboration' approach proposed in this paper?
It is a proposed way of running interdisciplinary AI research projects so that humanities/social-science (AHSS) partners and technical AI partners jointly define research problems, values and success criteria, instead of the common one-way pattern where the AHSS partner only supplies ethics input and the technical partner only builds a requested tool. It rests on reflexive practice (continual self-examination of one's disciplinary assumptions) and is expressed as five concrete practices.
What are the five key practices in the framework?
(1) Being open and upfront about jargon and disciplinary terms, since the same word can mean different things across fields; (2) interrogating how each discipline defines output, impact and novelty; (3) developing a common understanding of the most pressing research problems; (4) identifying values around AI use in context, going beyond simple technology-acceptance models; and (5) considering the role of individual habits and beliefs in sustaining collaborations.
Is the framework empirically tested?
No. This is a position paper: the authors derive the five practices from the literature on GLAM (galleries, libraries, archives, museums) convergence and from reflexivity theory, plus their own collaboration experience. They explicitly state the work is in progress and that future research will provide use cases and doctoral-training applications.
What does 'GLAM convergence' mean in this paper?
Since roughly the 1970s–80s, galleries, libraries, archives and museums have merged institutions, collections or professional standards — driven partly by digital technology — while their professions kept distinct theories and values. The paper cites both positive outcomes (higher cross-sector collaboration rates; improved mutual respect) and failures (convergence used for staff cuts, causing professional-identity backlash) as lessons for interdisciplinary AI collaboration.
About this review
This is an AI-generated preliminary review by Evolve.Pub, produced from the openly available text of the paper. It is not peer review and does not represent the authors' views. The canonical record of the paper is at the publisher link above.
- Figure 1 (barriers diagram, Section 2.2) is referenced but its visual content could not be assessed from the extracted text.
- Footnote bodies for in-text markers 2–3 are missing from the extraction.
- References never visibly cited in the body (Langlotz et al. 2024; Ryan et al. 2018; Schuitema et al. 2025) may be a truncation artifact rather than author error.
Go deeper on this paper
Wrote a paper? Stress-test it before reviewers do.
Upload your manuscript for the same structured scrutiny — reasoning, statistics, evidence, figures, and reproducibility.