AI's Impact on Culture: Navigating Opportunities and Challenges

Published Date : 19/06/2025 

Explore the potential of AI in transforming the cultural sector, from data analysis to content creation, while addressing the ethical and practical concerns. 

Is artificial intelligence (AI) a development that will help transform the cultural sector, or an existential threat that will destroy not only the arts but civilisation as we know it? Or something in-between the two?

In some senses, it is no surprise 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, and some of the more out-there possibilities are mind-blowing. But 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: “In my opinion, everyone can probably benefit from using AI in a sensible, measured way – and this includes museums.” And we should absolutely never forget that what we’re seeing now in AI is not actually anything like ‘intelligent’. They are just really great prediction engines. But this is not to say that the impact won’t be huge.

For Ellis, AI, in the same way as other technologies, 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. But he is also cautious. “If a curator was terrible and knew nothing, they could get AI to ‘write’ an exhibition – but it would be slop,” he says. “Whereas a good, already 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. A world in which websites or exhibitions are written entirely by AI will be a terrible world – because those exhibitions (like a wholly AI-generated website), by their nature, tend towards the most average thing in the world.

These tools don’t have opinions, they don’t have any edge and they don’t have an opinion or thoughts that are controversial – and those are all required for awesome journalistic, biting, exciting content.” 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.

“We are less interested in the generative side of it, which is what people are getting concerned about, and rightly so, because that feels problematic in all sorts of ways, particularly as museums are very trusted institutions,” says Gosling. “The interesting bit is the potential when you’ve brought together big datasets. You need ways of pulling together information that already exists about objects in order to make sense of this huge mass of data. That’s hugely exciting and we are actively hoping that is how the data is used.”

Whatever people’s concerns about AI, museums are using it 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. Transforming Collections is one of five Discovery Projects that formed the Towards a National Collection programme, which is funded by UKRI’s Arts and Humanities Research Council, and aims to open up collections to the public and researchers.

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. “This project was about working with the Decolonising Arts Institute to understand what art historians, collection owners and archivists need to do and what their challenges are, particularly with respect to understanding their collection,” he says. “We thought that once we had a good understanding of that, we could build a tool that provided something that actually did the job they want to do.

We tried to get to a point where people understood the technology well enough, so they can help us work out what they wanted to do with it – and that worked. We got to a stage by the end of the project where we have 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 research is complete, and we have a prototype, and we’re going to be showing that at a conference. We also have a workshop for a small number of people in the museum sector. The next step is to make it available, if we can.”

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 – including news articles, books, music, film, photography and visual art – 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. “There is a huge slab of thought that needs to go into the imbalance here,” he says. “It’s the same as any other power imbalance – but we need to think much more carefully about the fact that hyper-wealthy monster companies are sucking all our stuff into their databanks, and they’re not paying anyone anything for the pleasure. I don’t hold much hope that anyone will care enough to do anything about this – they didn’t and don’t when it comes to social media – but we really should be thinking hard about it from that point of view as well.” 

Frequently Asked Questions (FAQS): 

Q: What is the main concern about AI in the cultural sector?

A: The main concern is the potential for AI to be used to generate content that lacks depth, originality, and human touch, which could diminish the quality and uniqueness of cultural experiences.

Q: How can AI benefit museums?

A: AI can benefit museums by helping with data analysis, improving cataloging and data management, and augmenting the work of curators and researchers.

Q: What is the Museum Data Service?

A: The Museum Data Service is a collaborative project aimed at transforming how collections information is held, accessed, and shared across the UK's museums and collections.

Q: What is the Transforming Collections project?

A: The Transforming Collections project is a three-year initiative led by the University of the Arts London’s Decolonising Arts Institute to understand and enhance the use of AI in analyzing and reimagining art collections.

Q: What is the Make it Fair campaign?

A: The Make it Fair campaign is an effort by creative industry workers to highlight the unfair use of creative content by tech companies to train their AI models without proper controls, transparency, or payment. 

More Related Topics :