AI in the Museum: Connecting Futures

The Executable Web. When Museums Become Actionable by AI Agents

A Structural Shift in the Architecture of the Web A structural shift is unfolding beneath the surface of the web. For three decades, websites have been designed primarily for human navigation. Pages were structured visually. Interfaces were optimized for clarity and engagement. Search engines indexed content so that users could find it, read it and act upon it. Today, a new layer is emerging. It is not designed for human eyes, but for artificial agents. The idea is simple, yet transformative. Instead of forcing AI systems to imitate human behavior by clicking buttons and interpreting graphical interfaces, websites can expose structured functions directly. Search the archive. Register for an event. Book a ticket. Retrieve metadata. Access educational material. These actions can be declared as callable tools, accessible to machines through standardized protocols. The web would no longer be merely read by machines. It would be executed by them. From Interface Imitation to Structured Execution Until now, AI agents have had to “pretend” to be human users. They scrape pages, simulate cursor movements, interpret layouts and attempt to infer meaning from visual structures. This approach is fragile and inefficient. It depends on interface stability and often breaks when websites change. Emerging standards such as the Model Context Protocol and related experimental initiatives aim to move beyond this paradigm. They propose a structured interaction model in which publishers define the functions that agents are authorized to call, under controlled conditions. Instead of navigating through layers of design intended for humans, an agent can directly invoke a declared function and receive a structured response. This evolution does not signal the disappearance of graphical interfaces. Museums will continue to design websites for visitors. But alongside the human-facing layer, a second layer may develop. This new layer would expose services in a structured, machine-readable way. In practical terms, the web becomes two-layered. One layer serves human experience. The other serves machine execution. From Visibility to Actionability. The Rise of AEO In such an environment, the strategic question shifts. For years, institutions have focused on Search Engine Optimization. The objective was visibility. How do we rank. How do we appear in search results. How do we attract traffic. In an agent-mediated ecosystem, visibility alone may no longer be sufficient. The new challenge becomes actionability. Can an AI assistant directly query our collection database. Can it enroll a user in a workshop. Can it retrieve authoritative descriptions without distorting them. This shift is sometimes described as a move from SEO to Agent Engine Optimization, or AEO. The question is no longer only whether agents can find us, but whether they can act through us in a controlled and reliable manner. Why This Matters for Museums For museums, the implications are substantial. If AI assistants increasingly mediate access to information, they may become dominant gateways to cultural content. A visitor might ask an assistant to suggest exhibitions on climate change, retrieve primary sources on a historical event, or book a science workshop for a school group. If the museum’s digital infrastructure exposes structured services, the assistant can perform these actions directly and accurately. If it does not, the assistant will rely on secondary sources, approximations or aggregated data from platforms that do not belong to the institution. The stakes extend beyond convenience. They concern the circulation of knowledge. Museums are custodians of validated, curated and contextualized information. When AI agents become intermediaries, the integrity of that information depends on how it is accessed and transmitted. Structured exposure of functions allows institutions to define the rules of interaction. They can specify what is accessible, under what conditions, with what attribution requirements. They can embed source references, usage constraints and traceability mechanisms into the interaction layer itself. Governance, Sovereignty and Editorial Control This introduces a governance dimension. Who defines the callable functions? Which datasets are exposed? How is premium or subscription-based content protected? How is editorial value preserved? Archives, digitized collections and scholarly resources represent significant investments. If agents can retrieve and redistribute content without clear attribution or control, institutional authority may erode. Conversely, if museums define interoperable standards and maintain oversight of their machine-facing layer, they can strengthen their position within AI ecosystems. The discussion is therefore not purely technical. It touches on sovereignty. If large AI platforms define proprietary interaction standards, cultural institutions risk dependency. Their content could circulate primarily through external ecosystems, shaped by opaque algorithms. Open and standardized approaches to an executable web offer an alternative path. They enable institutions to remain actors rather than passive data providers. By declaring structured tools themselves, museums can retain agency in how their knowledge is mobilized. Rethinking Digital Strategy in an Agent-Mediated World This transformation also redefines metrics of digital success. Traffic and page views may no longer capture the full picture. A museum’s influence might increasingly be measured by how often its structured services are called by trusted agents, how frequently its authoritative descriptions are cited, and how effectively its systems integrate into educational or research workflows mediated by AI. Yet caution is necessary. An executable web amplifies both opportunity and risk. Poorly defined interfaces could expose sensitive data. Insufficient traceability mechanisms could weaken attribution. Over-automation could reduce the richness of human engagement. Museums must therefore approach this shift deliberately. The question is not whether to embrace or reject machine interaction, but how to design it in alignment with institutional values. At a practical level, cultural institutions may begin by identifying the services that could benefit from structured exposure. Collection search functions. Event registration systems. Educational resource access. Metadata retrieval. Each of these can be mapped and evaluated. Which actions should be callable. Which require authentication. Which must remain restricted. This mapping exercise is as much strategic as technical. Conclusion. Becoming Executable on Our Own Terms The emergence of an executable web marks a profound reconfiguration of digital infrastructure. It suggests that websites are no longer static presentations of information, but programmable knowledge systems. Museums have long adapted to technological change. From print catalogues to websites, from audio guides to immersive installations, they have integrated new media into their mission of transmission. The agent-driven web represents another stage in this evolution. It challenges institutions to rethink their digital architecture, not as a mere communication channel, but as an interoperable system within a broader computational ecosystem. The critical question is therefore not whether machines will execute parts of the web. That trajectory is already visible. The real question is whether museums will shape this layer proactively, or allow it to be shaped around them. In an agent-mediated world, remaining visible is no longer enough. Institutions must also remain executable, on their own terms. Benjamin BENITA [https://www.linkedin.com/in/benjamin-benita/], Editor-in-Chief, MuseumWeek Magazine Sources: Google previews WebMCP, a new protocol for AI agent interactions [https://searchengineland.com/google-releases-preview-of-webmcp-how-ai-agents-interact-with-websites-469024?utm_source=chatgpt.com] Agents no longer need to “pretend” to be human [https://eu.36kr.com/en/p/3678378253083529] Model Context Protocol (MCP) [https://github.com/modelcontextprotocol] Get full access to MuseumWeek Magazine at museumweek2h1r4.substack.com/subscribe [https://museumweek2h1r4.substack.com/subscribe?utm_medium=podcast&utm_campaign=CTA_4]

Recipes for Deploying AI Projects in a Museum

Generative AI and Museums. What Truly Creates Value for Audiences Generative artificial intelligence is gradually establishing itself in museums. Not only as a mediation tool, but also as a new way to explore collections and shape visitor experiences. Yet a central question remains for cultural institutions. What actually drives audiences to adopt these AI-powered tools. A recent academic study published in npj Heritage Science provides robust, data-driven answers, based on the analysis of The Living Museum, an experimental generative AI platform developed by the British Museum . A study grounded in real user experience The research draws on responses from 726 users, clearly distinguishing between cultural professionals and non-professional visitors.Its objective. To understand how perceived value is constructed and how it directly influences the intention to use generative AI in a museum context. The key factor. Relevance over spectacle The main finding is unequivocal.Audience adoption is not driven by the abstract promise of AI, but by two very concrete capabilities: * Semantic relevance. The ability of the AI to provide accurate, meaningful answers aligned with users’ questions and expectations. * Contextual adaptability. The capacity to adjust responses according to the visitor’s level of knowledge, intent (casual exploration or in-depth inquiry), language, and situational context. In other words, an AI perceived as accurate and well-situated creates more value than an AI designed primarily to impress. For museums, this reinforces a critical principle. AI must strengthen cultural authority, not undermine it. What increases perceived value Four factors significantly enhance perceived value: * Usefulness. Helping visitors understand, navigate, and explore collections more effectively. * Enjoyment. A smooth, engaging interaction that does not feel effortful. * Novelty. The feeling of discovering a new way to relate to heritage. * Relative advantage. Performing better than traditional tools such as labels, audio guides, or standard digital interfaces. By contrast, two elements clearly hinder adoption: * Perceived complexity, which disrupts immersion and generates cognitive fatigue. * Perceived risk, particularly regarding content reliability and data protection. One result is particularly striking. Explicit personalization does not significantly increase perceived value.In a museum context, visitors appear to prioritize scientific credibility and institutional trust over extensive configuration options. This insight has important implications for AI design in cultural settings. Perceived value drives adoption The study confirms a strong link between perceived value and intention to use.However, this relationship is moderated by two psychological factors: * Users with a high openness to innovation are more likely to translate positive experiences into sustained adoption. * Excessive interactivity can paradoxically weaken the impact of perceived value. When everything becomes interactive, clarity and depth may be lost. The message for museums is clear. More interaction is not always better. Balance matters. Professionals vs general audiences. Two distinct logics The research highlights a structural divergence between user groups: * Cultural professionals tend to value technological novelty and experimental potential. * General audiences are more sensitive to perceived risks and institutional guarantees. This implies differentiated strategies.A single AI system cannot be designed, framed, and deployed in the same way for all users. What this study changes for museums This research provides a clear framework for thinking about generative AI in museums: * AI is not primarily a technological issue, but a matter of cultural value perception. * Semantic accuracy, contextualization, and restrained interaction design are decisive. * Scientific authority and transparency become core design principles. * Strategies must be audience-specific, including at the interface level. For institutions engaged in MuseumWeek and beyond, this study serves as a valuable compass. It encourages museums to move beyond technological enthusiasm and toward a responsible, situated, and audience-centered approach to generative AI. Source: https://www.nature.com/articles/s40494-025-02194-9 [https://www.nature.com/articles/s40494-025-02194-9] Get full access to MuseumWeek Magazine at museumweek2h1r4.substack.com/subscribe [https://museumweek2h1r4.substack.com/subscribe?utm_medium=podcast&utm_campaign=CTA_4]

🎙️ Case Study – Archäologisches Museum Hamburg: "Photo Detective" and the Automated Analysis of Historical Archives

Introduction As part of our series exploring how AI is being implemented in museums worldwide, this case study focuses on the “Photo Detective” project at the Archäologisches Museum Hamburg (AMH). Led by Michael Merkel, this initiative tackles a common challenge for cultural institutions: managing vast collections of analogue photographs. While the AMH has digitized approximately 75% of its collection, the lack of detailed metadata makes these archives difficult for researchers and curators to navigate. Funded as a Proof of Concept by the InnotechHH Fund, “Photo Detective” uses automated tagging to transform these static images into a highly searchable digital resource. The Technological Dimension Object and Context Recognition The core of “Photo Detective” is an AI system driven by object recognition. The model was developed using a training set of 2,613 hand-annotated images. While the team considered 37 different object classes during the labeling phase, 21 classes were ultimately included in the final training model. The AI is capable of identifying a wide range of elements, from high-frequency subjects like “people” and “cars” to more specific architectural features like “timber framing,” “thatch roofs,” and “lattice windows.” Beyond identifying individual objects, the project explores context detection. For example, by recognizing specific clusters of objects—such as sports equipment or crowds—the AI can identify “sports events” as a general context. Remarkably, the technology’s potential extends beyond standard photography, successfully tagging historical engravings and postcards that include printed text. The “Human-in-the-Loop” Workflow A defining characteristic of this project is its “Human-in-the-Loop” centered workflow, which ensures that machine efficiency is balanced with human expertise. This six-step cycle creates a continuous loop of improvement: 1. Data Annotation: Humans manually label training data in a dedicated application. 2. Training: This data is used to teach the machine learning model. 3. Evaluation: Professionals assess the model’s performance to ensure the quality of predictions. 4. Hosting: The validated model is hosted on the “Photo Detective” platform. 5. Bulk Processing: Users initiate the automated tagging of large datasets. 6. Feedback: Ongoing user feedback is fed back into the database to refine future training phases. Impacts on the Cultural Sector The implementation of “Photo Detective” is reshaping several areas of museum practice: • Institutional Management: By automating the labeling process, the museum significantly reduces the administrative burden of archival processing, allowing staff to focus on high-level curation. • Knowledge Sharing: In a move toward collaborative innovation, the AMH plans to make its training data available as “open data” for other cultural institutions, helping the wider sector develop similar tools. • Research and Mediation: Enhanced searchability allows researchers to find specific historical details, such as every image featuring a “horse” or a “shop window”, instantly, opening new doors for historical analysis and public engagement. Perspectives and Issues The “Photo Detective” project highlights the shifting role of the museum professional in the age of AI. While the tool offers immense speed, the “Human-in-the-Loop” approach is essential to address the opacity of machine-generated interpretations. It ensures that the final tags remain accurate and contextually relevant. Additionally, as museums become more dependent on these digital tools, questions regarding the long-term sustainability of the technology and the standardization of data across different institutions remain at the forefront of the discussion. Conclusion The Archäologisches Museum Hamburg demonstrates how AI can revitalize historical archives. By combining automated object recognition with a rigorous human oversight process, “Photo Detective” makes history more accessible and participatory. This case serves as a model for how museums can use technology not just to store the past, but to make it a searchable and living resource for the 21st century. Links Archäologisches Museum Hamburg: https://amh.de [https://amh.de] Michael Merkel on LinkedIn: https://www.linkedin.com/in/michael-merkel-8759bb13/ [https://www.linkedin.com/in/michael-merkel-8759bb13/] Get full access to MuseumWeek Magazine at museumweek2h1r4.substack.com/subscribe [https://museumweek2h1r4.substack.com/subscribe?utm_medium=podcast&utm_campaign=CTA_4]

🎙️ How can a Cultural Project Manager use AI to anticipate risks?

You can listen to our podcast, read the full article, or watch the AI-generated video 📺 at the end of this page. Introduction – About the Podcast Welcome to the series Museum Professions: Working with AI, part of the AI in the Museum rubric by MuseumWeek. Each episode dives into a specific profession inside the museum world and explores how artificial intelligence is transforming daily practices. Today, we step into the shoes of a Cultural Project Manager. The Job and Its Challenges Cultural Project Managers oversee the planning, execution, and evaluation of projects within museums and cultural institutions. They are responsible for coordinating teams, managing budgets, and ensuring that projects meet organizational goals. Three major operational challenges include: 1. Unpredictable audience engagement, which can lead to underwhelming attendance and financial shortfalls. 2. Budget management complexities, where unexpected costs can derail project timelines and objectives. 3. Stakeholder alignment, as diverse interests can create conflicts that hinder project progress. Advertisement (internal promotion by MuseumWeek). How AI Can Help – Practical Solutions with Tools Challenge 1: Audience Engagement Forecasting The Problem: Anticipating audience engagement is a significant challenge for cultural project managers. Inaccurate predictions can lead to insufficient resources allocated for marketing and programming, ultimately affecting attendance and revenue. Understanding audience preferences and behaviors is essential for tailoring experiences that resonate with visitors. AI Approach: Tools such as predictive analytics and natural language processing (NLP) can analyze historical attendance data and social media sentiment to forecast audience engagement. By identifying patterns and trends, cultural project managers can make data-driven decisions to enhance outreach and engagement strategies. Implementation Path: A project manager could begin by collecting historical attendance data and social media interactions. Using a predictive analytics tool like Tableau [https://www.tableau.com], they can visualize trends and create forecasts. By integrating NLP tools to analyze audience sentiment from social media, they can refine their marketing strategies and tailor programming to better meet audience expectations. Risks & Limits: While AI can provide valuable insights, it is essential to consider potential biases in data and the ethical implications of using audience data. Additionally, reliance on AI tools may lead to overconfidence in predictions, which could result in inadequate contingency planning. Recommended tools: - Tableau [https://www.tableau.com] - Google Cloud AI [https://cloud.google.com/products/ai] Challenge 2: Budget Management The Problem: Managing budgets in cultural projects can be fraught with uncertainties, as unexpected costs can arise at any stage. In a museum context, this can lead to compromised project quality or even project cancellation, impacting overall institutional goals. AI Approach: AI-driven financial modeling and anomaly detection tools can help project managers identify potential budget overruns before they occur. By analyzing historical spending patterns and real-time data, these tools can alert managers to unusual spending behaviors, allowing for proactive adjustments. Implementation Path: A cultural project manager could utilize an AI financial modeling tool such as Adaptive Insights [https://www.adaptiveinsights.com] to create a dynamic budget model that adjusts based on real-time data. By integrating this tool with existing financial systems, they can monitor spending closely and receive alerts for any anomalies, ensuring that budget management remains on track. Risks & Limits: The reliance on AI for budget management introduces risks related to data accuracy and the potential for algorithmic bias. Furthermore, the cost of implementing advanced AI tools may be prohibitive for smaller institutions, necessitating careful consideration of resource allocation. Recommended tools: - Adaptive Insights [https://www.adaptiveinsights.com] - IBM Planning Analytics [https://www.ibm.com/cloud/planning-analytics] Challenge 3: Stakeholder Alignment The Problem: Ensuring alignment among diverse stakeholders can be a complex task for cultural project managers. Conflicting interests and priorities may lead to project delays or failures, undermining the collaborative spirit essential in cultural institutions. AI Approach: Collaborative AI tools and sentiment analysis can facilitate stakeholder communication and gauge the overall sentiment towards project goals. By analyzing feedback and discussions, project managers can identify areas of concern and address them proactively. Implementation Path: A project manager could implement a sentiment analysis tool like MonkeyLearn [https://monkeylearn.com] to analyze stakeholder feedback gathered through surveys and meetings. By categorizing sentiments and identifying key concerns, they can tailor their communication strategies to address issues and foster alignment among stakeholders. Risks & Limits: While AI can enhance stakeholder engagement, it is crucial to ensure that the data collected respects privacy and ethical considerations. Additionally, over-reliance on automated tools may overlook the nuances of human communication, potentially leading to misunderstandings. Recommended tools: - MonkeyLearn [https://monkeylearn.com] - Qualtrics [https://www.qualtrics.com] Looking Ahead – Tomorrow’s Possibilities In the next 12 to 24 months, cultural project managers can expect advancements in AI technologies to further enhance operational efficiency. Skills in data analysis and AI tool utilization will become increasingly important, necessitating training and development. However, governance frameworks will need to evolve to address ethical considerations and ensure equitable access to AI resources. Conclusion This episode has highlighted the potential of AI in anticipating risks faced by cultural project managers. By leveraging predictive analytics, financial modeling, and sentiment analysis, project managers can enhance decision-making and foster resilience in their projects. Reflective Questions: 1. How can your institution better integrate AI tools to improve risk management practices? 2. What ethical considerations should be prioritized when implementing AI in cultural project management? 3. How can collaboration among different departments enhance the effectiveness of AI applications in your projects? Get full access to MuseumWeek Magazine at museumweek2h1r4.substack.com/subscribe [https://museumweek2h1r4.substack.com/subscribe?utm_medium=podcast&utm_campaign=CTA_4]

🎙️ How Does AI Challenge the Role of Editorial Webmasters?

Introduction – About the Podcast Welcome to the series Museum Professions: Working with AI, part of the AI in the Museum rubric by MuseumWeek. Each episode dives into a specific profession inside the museum world and explores how artificial intelligence is transforming daily practices. Today, we step into the shoes of a Editorial webmaster. As the digital landscape evolves, editorial webmasters in museums face increasing pressure to manage content effectively while ensuring audience engagement. The integration of AI technologies presents both opportunities and challenges, as these professionals must navigate new tools and methodologies to enhance their workflows. Understanding the implications of AI on their roles is crucial for adapting to this rapidly changing environment. The Job and Its Challenges Editorial webmasters are responsible for curating, managing, and updating digital content across museum platforms. They ensure that information is accurate, accessible, and engaging for diverse audiences. Key challenges include: 1. Content Overload: The sheer volume of digital content can overwhelm webmasters, making it difficult to maintain quality and relevance. 2. Audience Engagement: Understanding and meeting the needs of varied audience segments is increasingly complex in a digital-first world. 3. Data Management: Efficiently managing and analyzing user data to inform content strategy poses significant challenges. How AI Can Help – Practical Solutions with Tools Challenge 1: Content Overload The Problem: Museums often produce vast amounts of digital content, leading to potential information overload for both webmasters and users. This can result in outdated or irrelevant content, diminishing the user experience and engagement. AI Approach: AI can streamline content management through natural language processing (NLP) and recommender systems. These technologies can help prioritize and curate content based on user preferences and behavior. Implementation Path: A museum professional could implement an NLP tool to analyze existing content and identify which pieces are most frequently accessed or shared. By using a recommender system like Google Cloud Recommendations AI [https://cloud.google.com/recommendations-ai], they can personalize content suggestions for users, enhancing engagement. The workflow would involve inputting user interaction data, processing it through the AI tool, and outputting tailored content recommendations. Risks & Limits: While AI can significantly enhance content management, it may introduce biases based on historical data. Additionally, the cost of implementing advanced AI tools can be prohibitive for some institutions, and governance around data privacy must be carefully managed. Challenge 2: Audience Engagement The Problem: Engaging diverse audiences requires a nuanced understanding of their interests and preferences, which can be difficult to ascertain without proper tools. Museums risk alienating segments of their audience if they fail to tailor content effectively. AI Approach: AI-driven analytics tools and chatbots can help gather insights on audience behavior and preferences. By leveraging these technologies, webmasters can create more targeted and engaging content. Implementation Path: A museum could deploy a chatbot, such as Dialogflow [https://dialogflow.cloud.google.com/], to interact with visitors on their website. This chatbot can collect data on user inquiries and preferences, which can then be analyzed to inform content strategy. The workflow would involve setting up the chatbot, integrating it with the website, and using the collected data to refine content offerings. Risks & Limits: The use of chatbots raises concerns about user privacy and data security. Additionally, reliance on AI for audience engagement may overlook the importance of human interaction, which can be vital in a museum context. Challenge 3: Data Management The Problem: Managing and analyzing user data effectively is crucial for informing content strategy, yet many webmasters lack the tools to do so efficiently. Poor data management can lead to missed opportunities for audience engagement. AI Approach: Machine learning algorithms and data visualization tools can assist in processing large datasets and extracting actionable insights. These technologies can help webmasters make informed decisions based on user interactions. Implementation Path: A museum could utilize a data visualization tool like Tableau [https://www.tableau.com/] to analyze user engagement metrics. By importing data from various sources, webmasters can create interactive dashboards that highlight trends and patterns in audience behavior. This enables them to adjust content strategies accordingly. Risks & Limits: The complexity of data governance can pose challenges, particularly regarding compliance with privacy regulations. Additionally, there is a risk of misinterpreting data without proper expertise, which could lead to misguided content decisions. Looking Ahead – Tomorrow’s Possibilities In the next 12 to 24 months, the role of editorial webmasters will likely evolve as AI tools become more integrated into museum operations. Professionals will need to develop new skills in data analysis and AI tool management while ensuring ethical governance of user data. Opportunities for enhanced audience engagement and personalized content will grow, but institutions must remain vigilant about the potential constraints of technology reliance. Conclusion This episode highlights the transformative impact of AI on the role of editorial webmasters in museums. By addressing challenges such as content overload, audience engagement, and data management, professionals can leverage AI to enhance their workflows and improve user experiences. Reflective questions for museum teams include: - How can we balance the use of AI tools with the need for human oversight in content management? - What strategies can we implement to ensure ethical governance of user data while utilizing AI technologies? Get full access to MuseumWeek Magazine at museumweek2h1r4.substack.com/subscribe [https://museumweek2h1r4.substack.com/subscribe?utm_medium=podcast&utm_campaign=CTA_4]