Theme Watch · a living map of AI across museums, galleries and cultural institutions · monthly scan · as at 11 Jul 2026.
Why we watch this theme. Museums are at a funding and attention inflection point, and AI is arriving as both a lifeline and a risk: it can reanimate a collection into a conversation, or flatten it into what critics call art slop. Dominic cares about the museum experience as a thing in itself, so this tracker follows where AI genuinely deepens a visit and where it hollows it out, who is building the tools, which institutions are leading, and the harder questions of trust, data ownership and human judgement underneath. Seeded by the FT piece below, then expanded by monthly scan (every new source is approved before it is added).
Current map: the live sub-threads
Conversational visitor experiences. Chatbots, reanimated historical figures, and "talk to the object" guides (Leighton House, the Met, the Dalí Museum, Ask Mona, Artlas).
Trust, accuracy and the "AI slop" critique. Hallucination and cultural homogenisation set against the museum's role as a trusted source.
Bespoke curated models vs generic LLMs. "Garbage in, garbage out": institution-curated datasets and museum-approved content versus open web models.
Museum data as an asset. Licensing collections and specialist data to AI developers (the UK Creative Content Exchange pilot; the Natural History Museum's biodiversity-data deal with BlackRock).
The funding crisis driving adoption. Falling visitor numbers, slashed public funding and an ageing donor base pushing museums towards AI.
Build vs buy, the "intelligent customer". In-house AI teams versus vendor platforms, and knowing which to choose.
Agent-readable museums. The "executable web": exposing collections so AI agents can act on them, and the data-sovereignty questions that follow.
Ethics, environment and human friction. Preserving creativity and critical thinking, and the energy cost of AI in culture.
Organisations we track
Vendors and platforms
Ask Mona· French conversational-AI firm (Marion Carré, 2017); museum chatbots and guides at the Louvre, the Pompidou and Versailles.
Artlas· AI museum-companion app; camera artwork recognition and personalised multilingual guides (Mori Art Museum, ICA Miami).
Google Arts & Culture· Google's platform putting museum collections online; an AI and digitisation partner to thousands of institutions.
Institutions
Natural History Museum, London· In-house AI team for over five years; licensed biodiversity data to BlackRock; part of the UK data-marketplace pilot.
Metropolitan Museum of Art· Natalie Potter chatbot for the 2024 Costume Institute; founding partner of the Museums + AI Network.
Smithsonian Institution· Its AI Values Statement drew on the Museums + AI Network toolkit; large AI and digital programmes.
Networks
The Museums + AI Network· AHRC-funded network (Oonagh Murphy, Elena Villaespesa) on AI, ethics and museums, with a widely-translated planning toolkit.
People we track
Marion Carré· Co-founder & CEO, Ask Mona. Argues for bespoke, curated models and for keeping productive "friction" in the experience. Teaches art and AI at the Sorbonne.
András Szántó· Author and museum strategist; adviser to the Met and to Artlas. Frames the future of museums as an exercise in leadership and values.
Doug Gurr· Director, Natural History Museum London (ex-Amazon; now also CMA chair). Leads the data-monetisation and "intelligent customer" argument.
Nataliya Kosmyna· Research scientist, MIT Media Lab (Fluid Interfaces). The critic voice: "AI art slop", cultural homogenisation, over-reliance harms.
Anders Sundnes Løvlie· Associate Professor, IT University of Copenhagen. Research on museum chatbots and the unreliability of LLMs in an authority context.
Oonagh Murphy· Senior Lecturer, Goldsmiths; co-founder of the Museums + AI Network. AI ethics and museum planning.
Seb Chan· Director & CEO, ACMI; previously digital lead at Cooper Hewitt (Smithsonian) and the Powerhouse. A leading museum-digital thinker.
Recent media
Verified articles, podcasts and talks on AI and museums, most recent first. As at 11 Jul 2026.
Museums to take part in AI marketplace pilot· Museums Association · 3 Feb 2026 Twelve UK cultural organisations join the government's Creative Content Exchange, letting AI developers access digitised assets, with creator-body warnings about unauthorised use.
Beyond the Scroll: Museums, AI, and the Value of Attention· The Harvard Crimson · 29 Dec 2025 Weighs authorship and labour concerns around AI-generated art, and argues museums can curate exhibitions that prompt reflection rather than replace human creativity.
The Executable Web: when museums become actionable by AI agents· AI in the Museum: Connecting Futures · 25 Feb 2026 Why museum sites must expose machine-readable interfaces for AI agents, the shift from search to "agent engine optimisation", and editorial control.
Where does computer vision stand for museums?· AI in the Museum: Connecting Futures · 12 Sep 2025 Practical computer-vision uses in museums, including accessibility apps, artwork authentication and exhibition optimisation.
FARI Brussels Conference 2025: Marion Carré keynote· FARI, AI for the Common Good Institute · Dec 2025 Ask Mona's latest work in culture, including an initiative at Versailles, and reframing the question from what AI can do for art to what art can do for AI.
Monthly changelog
11 Jul 2026 · Theme Watch, AI Museums launched. Baseline built from the FT seed article plus verified anchors: 13 organisations, 7 people, and recent media across articles, podcasts and talks. No change markers yet (first baseline; markers begin next cycle).