{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,30]],"date-time":"2026-10-30T17:38:13Z","timestamp":1793381893815,"version":"4.5.0"},"reference-count":19,"publisher":"CERN","license":[{"start":{"date-parts":[[2026,7,28]],"date-time":"2026-07-28T00:00:00Z","timestamp":1785196800000},"content-version":"unspecified","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"name":"European Commission","award":["101136910"],"award-info":[{"award-number":["101136910"]}],"id":[{"id":"https:\/\/ror.org\/00k4n6c32","id-type":"ROR","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Open Res Europe"],"abstract":"<jats:p>Background\n                    Smallholder farmers on remote, depopulating islands have limited access to advisory services, and their locally specific agronomic knowledge, often expressed in regional dialect, is poorly represented in the corpora on which large language models are trained. A general-purpose chatbot therefore answers local questions fluently but unreliably. A conversational artificial intelligence assistant, \n                        Falco eleonorae, was developed for the farmers and cooperatives of Kythera and Antikythera, Greece.\n                \n                \n                    Methods\n                    The assistant is a thin Backend-for-Frontend proxy in front of a managed, geospatially aware agronomic agent, hosting no chat model of its own. Answer generation and tool selection are delegated to upstream models; a vision-capable model turns an uploaded field photograph into a short description, so only text reaches the agent; voice is transcribed by a managed speech-to-text service in a European Union region. Grounding uses tool-augmented retrieval rather than a self-hosted vector database: a Model Context Protocol tool queries a curated, read-only, bilingual data interface, and every request carries a geospatial Well-Known Text envelope anchoring the agent to the islands.\n                \n                \n                    Results\n                    It is deployed as a Greek-primary progressive web application supporting text, voice, and image input with streamed responses, suited to low digital literacy and poor connectivity. Locally scoped questions are answered from the pilot\u2019s editorially controlled records rather than the model\u2019s parametric memory, and content edits take effect on the next query without re-indexing. Safeguards include European Union data residency for speech, server-side editorial gating, rate limiting, output sanitisation, and erasure of chat content on account deletion.\n                \n                \n                    Conclusions\n                    For a small, resource-constrained rural deployment, a managed and grounded multilingual assistant is more attainable and trustworthy than a bespoke self-hosted model. No formal user evaluation has yet been performed, and a structured field evaluation is the principal future work.<\/jats:p>","DOI":"10.12688\/openreseurope.24512.1","type":"journal-article","created":{"date-parts":[[2026,7,28]],"date-time":"2026-07-28T06:29:09Z","timestamp":1785220149000},"update-policy":"https:\/\/doi.org\/10.68336\/openreseurope.updatepolicy","source":"Crossref","is-referenced-by-count":1,"title":["Retrieval-grounded multilingual LLM assistance for island smallholder farmers"],"prefix":"10.68336","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-6727-1014","authenticated-orcid":true,"given":"Nikolaos D.","family":"Tantaroudas","sequence":"first","affiliation":[{"name":"Ethniko Metsobio Polytechneio Ereunetiko Panepistemiako Institouto Systematon Epikoinonion kai Ypologiston, Athens, Attica, 15773, Greece"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ilias","family":"Karachalios","sequence":"additional","affiliation":[{"name":"University of Thessaly, Department of Environment, School of Technology, Larissa, 41500, Greece"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Andrew J.","family":"McCracken","sequence":"additional","affiliation":[{"name":"DASKALOS-APPS, PERONNAS, 01960, France"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"13904","published-online":{"date-parts":[[2026,7,28]]},"reference":[{"key":"116896","doi-asserted-by":"publisher","DOI":"10.3390\/su16167068"},{"key":"116897","doi-asserted-by":"publisher","DOI":"10.1038\/s43016-023-00867-x"},{"key":"116898","doi-asserted-by":"publisher","DOI":"10.1177\/24551333251345224"},{"key":"116899","unstructured":"4. PoliRURAL Plus Consortium: PoliRURAL Plus: Building capacity for foresight-driven rural innovation (Horizon Europe). 2024. Reference Source (https:\/\/poliruralplus.eu\/)"},{"key":"116900","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2606.25647"},{"key":"116901","unstructured":"6. Vaswani A, Shazeer N, Parmar N, et al.: Attention is all you need. Advances in Neural Information Processing Systems (NeurIPS). 2017; 30: 5998\u20136008. Reference Source (https:\/\/proceedings.neurips.cc\/paper\/2017\/hash\/3f5ee243547dee91fbd053c1c4a845aa-Abstract.html)"},{"key":"116902","unstructured":"7. Brown TB, Mann B, Ryder N, et al.: Language models are few-shot learners. Advances in Neural Information Processing Systems (NeurIPS). 2020; 33: 1877\u20131901. Reference Source (https:\/\/proceedings.neurips.cc\/paper\/2020\/hash\/1457c0d6bfcb4967418bfb8ac142f64a-Abstract.html)"},{"key":"116903","doi-asserted-by":"crossref","unstructured":"8. Ouyang L, Wu J, Jiang X, et al.: Training language models to follow instructions with human feedback. Advances in Neural Information Processing Systems (NeurIPS). 2022; 35: 27730\u201327744. Reference Source (https:\/\/proceedings.neurips.cc\/paper\/2022\/hash\/b1efde53be364a73914f58805a001731-Abstract-Conference.html)","DOI":"10.52202\/068431-2011"},{"key":"116904","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2212.08073"},{"key":"116905","unstructured":"10. OpenAI: GPT-5 system card. Technical report. OpenAI;2025. Reference Source (https:\/\/cdn.openai.com\/gpt-5-system-card.pdf)"},{"key":"116906","unstructured":"11. Anthropic: System card: Claude Sonnet 4.5. Technical report. Anthropic;2025. Reference Source (https:\/\/www.anthropic.com\/claude-sonnet-4-5-system-card)"},{"key":"116907","unstructured":"12. Lewis P, Perez E, Piktus A, et al.: Retrieval-augmented generation for knowledge-intensive NLP tasks. Advances in Neural Information Processing Systems (NeurIPS). 2020; 33: 9459\u20139474. Reference Source (https:\/\/proceedings.neurips.cc\/paper\/2020\/hash\/6b493230205f780e1bc26945df7481e5-Abstract.html)"},{"key":"116908","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2020.emnlp-main.550"},{"key":"116909","unstructured":"14. Yao S, Zhao J, Yu D, et al.: ReAct: Synergizing reasoning and acting in language models. International Conference on Learning Representations (ICLR). 2023. Reference Source (https:\/\/openreview.net\/forum?id=WE_vluYUL-X)"},{"key":"116910","unstructured":"15. Anthropic and Model Context Protocol Contributors: Model Context Protocol specification (revision 2025-11-25). 2025. Reference Source (https:\/\/modelcontextprotocol.io\/specification\/2025-11-25)"},{"key":"116911","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2604.03057"},{"key":"116912","doi-asserted-by":"publisher","DOI":"10.12688\/openreseurope.23201.1"},{"key":"116913","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2604.06901"},{"key":"116914","doi-asserted-by":"publisher","DOI":"10.34190\/iccws.21.1.4465"}],"container-title":["Open Research Europe"],"original-title":[],"link":[{"URL":"https:\/\/open-research-europe.ec.europa.eu\/articles\/6-262\/v1\/xml","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/ore.eu\/ore\/article\/download\/1457\/version\/2425\/4847","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/ore.eu\/ore\/article\/download\/1457\/version\/2425\/4848","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/ore.eu\/ore\/article\/download\/1457\/version\/2425\/4847","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,30]],"date-time":"2026-10-30T17:14:21Z","timestamp":1793380461000},"score":1,"resource":{"primary":{"URL":"https:\/\/ore.eu\/ore\/article\/view\/1457\/version\/2425"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,7,28]]},"references-count":19,"URL":"https:\/\/doi.org\/10.12688\/openreseurope.24512.1","relation":{"has-review":[{"id-type":"doi","id":"10.21956\/openreseurope.26557.r81383","asserted-by":"object"},{"id-type":"doi","id":"10.21956\/openreseurope.26557.r79066","asserted-by":"object"},{"id-type":"doi","id":"10.21956\/openreseurope.26557.r80040","asserted-by":"object"}]},"ISSN":["2732-5121"],"issn-type":[{"value":"2732-5121","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,7,28]]},"assertion":[{"value":"2026-07-28","name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}},{"value":"This project has received funding from the European Union's Horizon Europe research and innovation programme through the PoliRURAL Plus project (Grant Agreement No. 101136910), via its DEVELOP cascade-funding open call. Views and opinions expressed are those of the authors only and do not necessarily reflect those of the European Union or the granting authority. Neither the European Union nor the granting authority can be held responsible for them.","name":"funding_statement","label":"Funding Statement","group":{"name":"funding_information","label":"Funding Information"}}],"article-number":"262"}}