AI Pros and Cons for Museums

Is artificial intelligence (AI) a development that will help transform the cultural sector, or an existential threat? There is alarm and confusion over the potential impact of AI. Recent developments seem to have happened very quickly, even for the rapidly paced tech world. This is even more reason to try to look at the future of AI with a clear head, to give museums the confidence to take advantage of the opportunities while being aware of the pitfalls.

Mike Ellis, the director of consultancy Thirty8 Digital, says everyone can probably benefit from using AI in a sensible, measured way, including museums. He adds that what we’re seeing now in AI is not actually anything like ‘intelligent’ and that they are just really great prediction engines. For Ellis, AI, will be useful for some things and not for others. He is already using it to help with coding and can see the benefits of deploying it to do certain things with large amounts of data. He says that a good, knowledgeable curator could use AI to augment their knowledge, bring new ideas, frame stuff in different ways and use AI to do a lot of the grunt work.

Ellis believes a world in which websites or exhibitions are written entirely by AI will be a terrible world because those exhibitions, by their nature, tend towards the most average thing in the world and don’t have opinions or thoughts that are controversial which are required for awesome content.

AI for Data Analysis

Many people working in museums are excited about the possibilities of AI helping to analyse data, but in the same way as Ellis, they are wary of using it to generate original content. Kevin Gosling, the outgoing director of the Collections Trust, thinks AI will be very useful for the Museum Data Service, which was launched last year to transform how collections information is held, accessed and shared across the sector. This collaboration between Art UK, the Collections Trust and the University of Leicester is aiming to bring together more than 100 million museum records from 1,750 Accredited museums and other collections across the UK.

Gosling says they are less interested in the generative side of it, which is what people are getting concerned about. He says the interesting bit is the potential when you’ve brought together big datasets and need ways of pulling together information that already exists about objects in order to make sense of this huge mass of data.

AI Projects in Museums

Museums are using AI already and the sector is actively exploring how it can work best, including the use of large datasets. One such project is Transforming Collections: Reimagining Art, Nation and Heritage, a three-year initiative led by Susan Pui San, a professor of contemporary art and the director of the University of the Arts London’s (UAL) Decolonising Arts Institute (DAI). The research project, which ran from November 2021 to January 2025, was carried out by colleagues from the DAI and UAL’s Creative Computing Institute, working closely with Tate and a further 14 national and international collections and archives.

Mick Grierson, a research reader at the University of the Arts London’s Creative Computing Institute, was one of the co-investigators on the Transforming Collections: Reimagining Art, Nation and Heritage project. He says the project was about understanding what art historians, collection owners and archivists need to do and what their challenges are, particularly with respect to understanding their collection. He adds that they tried to get to a point where people understood the technology well enough, so they can help work out what they wanted to do with it. By the end of the project they had a tool you can log into that lets you pick a dataset you’re working on, and you can build your own machine learning models to investigate that collection using examples you provide. The next step is to make it available, if they can.

Big Tech Concerns

Hanging over any discussion of AI is the power of big tech and the threat it poses to how people and institutions can control their creative output. The recently launched Make it Fair campaign is an effort by those working in the creative industries to highlight how tech companies using creative content train their generative AI models. Publishers and creators say that doing this without proper controls, transparency or reasonable payment is unfair and threatens their livelihoods. They are also concerned that the UK government wants to change the laws to favour tech platforms, so they can use creative content to power their AI models without permission or payment unless the creators specifically say “no”.

This power of big tech is certainly a concern for Ellis, who says that hyper-wealthy monster companies are sucking all our stuff into their databanks, and they’re not paying anyone anything for the pleasure.