Technology
The investigation progresses to the group of questions on the role of technology in accessing and sharing knowledge artefacts. Firstly, the hosting of knowledge re-sources by the two most-mentioned websites (from the interviews and the online questionnaires) are investigated; CultureHive, and The Collections Trust.
The CultureHive is a well-maintained and visually pleasant, uncluttered website managed by the Arts Marketing Association (AMA) (AMA, 2022a, 2025b; AMA CultureHive, 2022). It is a ‘free online resource hub for cultural professionals that brings the collective intelligence of the sector together in one place, by you, for you’. CultureHive is a host website, rather than a creator. The knowledge artefacts hosted in it are not only freely available, but also have the requirement to follow the Heritage Lottery guidance on creative commons open licence use (pers. comm. CultureHive). As shown by the data, its extensive content is regularly used by participants, but was of-ten perceived as not broad enough due to marketing being its main objective. The CultureHive landing webpage layout focused on highlighting the latest and most-read resources. This is one approach to dealing with the problem of quickly outdated knowledge resources, which is a significant issue as seen in the analysis of this re-search. The full range of knowledge resources was findable through the search func-tion leading to a separate webpage. This webpage provided a search function by re-search type, by smart tags, and by learning blocks. These are all very useful func-tions, and the learning blocks in particular connect to professional development and also to the same thinking behind the AIMs hallmarks; they are mainly overarching concepts such as audience behaviour, or outreach and collaboration. These were in-fluential at the start of this research in devising the operational framework; the Cul-tureHive learning blocks were developed based on four pillars which ‘… help identify the skills and knowledge needed to achieve your objectives’ (Figure 7 25) (AMA, 2022b) and the operational framework in this research is designed informed by this approach.
The CultureHive had another, understated but important webpage, the Service Provider Hub (AMA, 2025c). As seen from the earlier data analysis, provider organi-sations have limited capacity and great need for knowledge too, on how to create and disseminate knowledge resources. Alongside a supportive digital infrastructure for providers, a knowledge curating design needs to be catering for knowledge needs of provider organisations, as much as it caters for user organisations.
CultureHive was initially set up, and has been periodically supported, through grant funding without which no major updates or improvement can take place. As such, it is also exploring the role of AI for the creation of, and signposting to, knowledge resources (pers. comm. AMA).
The Collections Trust manages potentially the most visited website in the eco-system, judging by the participant comments and questionnaire contributions. It was praised for its useful, authoritative, and trusted content, and for being well-maintained with no broken links. The Collections Trust provides training, professional services, and manages ‘Spectrum’, the core collection management standard around the world, a structured guidance comprising 21 procedures (Collections Trust, 2025a). The Collections Trust as an Investment Principles Support Organisation (IPSO) is the Art Council’s right arm and depository for everything to do with collections management in relation to accreditation. Its knowledge artefacts were organised extremely well and presented in a clear, attractive, and consistent website architecture (Collections Trust, 2025c). To accompany the knowledge artefacts, the Collections Trust provides train-ing; the collections documentation training course attended by the author in 2024, was delivered by the Collections Trust confidently, run like a clockwork, was informa-tive and was followed promptly with copies of the slides and accompanying links. Fur-thermore, in 2024 the Museum Data Service was launched as a joint initiative by the Collections Trust, Art UK, and the University of Leicester, supported by Bloomberg Philanthropies and the Arts and Humanities Research Council (MDS, 2025). This knowledge curating research was originally influenced, on operational knowledge ra-ther than curatorial, by the early aspiration of the Museum Data Service project; for museums ‘to tap the full potential of data they already have, through work they are already doing, and with funding streams that – mostly – already exist’ (Collections Trust, 2022b).
The Collections Trust does not seem to churn out high volumes of resources like some provider organisations do, but to create few core resources that have lon-gevity and quality. Over time the Collections Trust has addressed the full spread of operations in its remit, which sits in the tangible objectives of the operational frame-work. As seen in the data analysis, its knowledge resources positively skew upwards the presence levels of knowledge artefacts for the entire Environmental preservation objective. In parallel, the resources hosted by CultureHive relate mainly to the intan-gible objectives (Cultural education and Social participation). One cannot but wonder then, the potential benefits of a closer partnership and sharing of expertise between these two key platforms and their organisations.
The mentioned websites host internally and externally created knowledge arte-facts, and signpost to other websites. If a central website was to be used to coordi-nate knowledge resources, then its branding, design, and content would require sig-nificant planning (alongside costs and maintenance). Several Knowledge Manage-ment Systems (KMS) and Customer Manager Systems (CRM) are available. An eco-system-wide platform would require a combination of KMS and CRM system applica-tions and further personalisation with several add-on apps for the different functions required (pers. comms, Charity Digital). One good knowledge base software example is KnowledgeOwl, with a user-friendly interface and many useful functions included (KnowledgeOwl, 2024). Considerations for a sector-wide platform include security (login access) in a way that does not become a barrier to users, data capacity, seam-less app integration, and user-friendly functionality (such as search filtering, adminis-trator levels, publishing, archiving, and version management) (pers. comm. Charity Digital, pers. comm. KnowledgeOwl). As a knowledge resources package that would link to training and development, it would also need integration with a Learning Man-agement System (LMS), with several systems being available, offering a range of functions at a variety of costs (Knowledge Integration, 2025; Omnia, 2024; Whale, 2024). Furthermore, the inductive analysis showed that lack of an explicit open li-cence was a barrier for organisations to adopt and adapt knowledge artefacts. Open access therefore is a key consideration and there is advice available specifically for cultural heritage by Creative Commons Open Culture (Creative Commons, 2025) and the GLAM e-lab which ‘… works directly with GLAM institutions to develop open ac-cess solutions accessible to the wider community of Galleries, Libraries, Archives, and Museums’ (GLAM-E Lab, 2024). Open access also relates to knowledge contri-bution by the whole ecosystem, calling for a Wiki-type function. Wiki sites are collabo-rative technologies in which the content is maintained by its users (Brainin and Arazy, 2016; Britannica Encyclopaedia, 2024b). Wikipedia is the most well-known and there is wiki software provided by a number of software companies, including an integrated wiki system within Microsoft packages (Microsoft, 2024). Wikimedia provides support and guidance for setting up and maintaining wiki pages, and a Wikibase cloud can be set up specifically for cultural heritage, which can also link to existing Wikipedia pages ((pers. comm. WikimediaUK), (The Wikibase Consultancy, 2024)). KMSs, CRMs, LMSs, wikis, and all associated apps are increasingly equipped with Artificial Intelli-gence (AI) capabilities.
The use of Artificial Intelligence (AI) is a key consideration for the knowledge curating design. Organisations are exploring its use to improve (or replace) searching, and for provider and user organisations to create knowledge artefacts (Analytics En-gines, 2025). The National Archives researched the AI’s capability for categorising and finding documents. The conclusion showed:
‘… promising results were obtained overall with no tool or approach consistently outperforming the others across all tasks … commercially available AI tools and pipelines can be successfully applied to aid the task of records selection in semi-structured and unstructured collections’ (Ranade, 2021; The National Archives, 2021).
The Stonnington Libraries in Australia utilised Microsoft AI to create its searchable catalogue. The project involved the development of searchable metadata for every record (ZDNET, 2019). This prompts the consideration of a process within the knowledge curating design for including suitable AI search metadata on all the knowledge resources being produced.
A key current AI research project is HAZEL, piloting a generative AI guidance assistant for Historic England. HAZEL is an ‘advice finder’ which ‘… will provide text summarisation and analysis capabilities to support those producing advice docu-ments, offering suggestions for enhancing and extending content’. The quality of the AI outputs from the different service providers varies significantly, and some are less suited to the cultural heritage content (pers. comm. HAZEL project). The HAZEL re-searchers and many of this research’s participating organisations are aware of the AI hallucinations, which are ‘… incorrect or misleading results that AI models generate’ (Google Cloud, 2025). One example of a hallucination is that ‘it recommends books that don’t exist’ (Toews, 2023). Work on reducing AI hallucinations, and improving the overall AI output is progressing at speed and its role is expected to be increasingly significant within the next few years ((pers. comm. AMACultureHive, pers. comm. HAZEL project), (University of Oxford, 2024)). The quality of the AI output depends highly on training the programme on relevant, high quality, trusted sources. Other-wise, the output does not only include hallucinations, but also reflects existing preju-dices in literature, for example reinforcing racism (Noble, 2018). As part of this re-search’s knowledge curating design, it is important to plan the development of a main-tained library of trusted resources (with their metadata), organised across the opera-tional framework, that will be the training stock for the application of AI in cultural heritage.
Relating training to learning styles, the data analysis showed that a wide range of knowledge artefact media are needed. Not just physical or just online documents, but also interactive forms, videos, and easily updated online how-to-guides. This cre-ates a requirement for provider organisations to have access to a much wider range of digital skills, associated software, and budgets. Examples of available software programmes include Vimeo for creating and hosting short introduction and longer training videos, or Scribe for recording the process steps of actions as a more visual alternative to a written operating procedure (Scribe, 2025; Vimeo, 2024). A number of systems convert text to video, in 140 languages, unsettlingly being presented by AI-created presenters (avatars) (Descript, 2025; HumanPal, 2025; Synthesia, 2025). Alongside the provider needs for creating diverse media of knowledge artefacts, a need was raised in the inductive analysis for user organisations to more easily and habitually record experiential knowledge, while an activity progresses. Knowledge then can be utilised more quickly and is not lost when people leave. There are a number of Electronic research notebooks (Wikipedia ELNs, 2024), simple versions of which could have application potential in creating a good habit of regularly recording steps in activities, in addition to process-recording systems like Scribe that was men-tioned earlier. AI is now integral part of all these applications, and AI is directly ac-cessed within Microsoft’s Word, Excell, PowerPoint, and Teams (Copilot), and Google’s equivalent apps (Gemini), simplifying some of the more administrative and tedious manual operations (Google Workspace, 2025; Microsoft Copilot, 2025). There are a lot of procedural and ethical questions related to these applications, as well as legal implications, such as consequences of developing a policy directly or indirectly informed by AI hallucinations or preexisting prejudice.
The technology changes quickly and a knowledge curating design ought to plan for the provider organisations to be fully trained and equipped, so they can utilise digital technology at ecosystem scale, and to be able to offer advice on tested systems. This would be more sensible than expecting thousands of small individual organisations to experiment with the variable quality of fast changing apps, and the as-sociated costs. In a knowledge curating design, the provider organisations need to be the leaders in the use and the development of digital technology for cultural heritage. One example for need in better utilisation of technology by providers is the accreditation’s Collections development policy template. The Arts Council has developed a very useful and compact template that all museums can use. This is still provided as a Microsoft Word (or Google Docs) document, which the user organisations must fill in or update manually every five years for the [re]accreditation. Users have to manu-ally amend sections depending on being in England, Scotland, Wales, or Northern Ire-land, and whether they are a government-funded national museum or not (ACE, 2018b).
Its administrators collate the completed forms manually from emails and then, presumably, update forms to track who has completed them and when a review is due. All this can be automated, with the documents being submitted online through a Microsoft or Google type Forms, which will be able to summarise responses and provide data for the whole ecosystem instantly. Most content, including the museum details and selection of all relevant text for the corresponding organisation could be auto filled by simply selecting the accreditation number (or working-towards-accreditation number). User organisations would then be able to make future updates online and the responses would update in real time, providing central data at ecosystem scale. The technology, however, is still bitty; Microsoft Forms, and Google Forms, do not inte-grate automatically with their own Word or Google docs respectively, without using other programmes, code, or upgraded paid software. Even then, without additional programming, there is no facility for the users to have a direct and readily access to the document they submitted through Forms. The author has discussed with ACE the development of a pilot automation system for this template (pers. comm. ACE ac-creditation). Additionally, each of the different collection templates would need to be manually developed and coded, which makes the automation of all the templates dif-ficult to set up and maintain, and costly. The technology, even with AI, is simply not user-friendly yet, it is driven by the software developers’ commercial strategies, not the user needs and abilities (Jones, 2024; TechTarget, 2024; Zhang, 2022). Overall, abductive logic suggests that significant digital investment needs to be focused on the provider organisations, as this will have a positive impact at large scale, across the whole ecosystem. Such approach could save time, hustle, and stress, across the thousands of user organisations.
