Museum Uses Artificial Intelligence to Curate Better Exhibitions

Artificial intelligence (AI) may have been around for some time but museums are only just getting used to the sort of functions it can perform. That said, there have been some impressive uses of this type of technology in the museum sector before where learning algorithms, in particular, have been deployed to predict human behaviours. However, Bologna’s Civic Museum has surprised many by making use of AI in an entirely new way – to help improve its curatorial prowess. Of course, what constitutes a well-curated exhibition is a matter of taste. Some might say that it is just a question of personal preference, almost like trying to decide what constitutes good art. Why then, should a museum see fit to use AI rather than human intelligence to curate its shows?

The idea driving the project is to give museum attendees more of what they like and less of what they don’t. In this regard, the principle behind the use of the technology is much the same as a museum or gallery director might demand of their flesh and blood curators. What most museums and public galleries are after, in the end, is augmented visitor numbers, after all. And yet, is it possible that AI can be better than skilled curators with decades of experience? Could technology really put together a better exhibition that boosts attendance figures? More intriguingly, perhaps, is the question of whether the shows produced in this manner will be better or simply more popular. In this regard, we return to the question of personal preference and taste, of course. So, what is going on?

Emotional Reactions

In June, an Italian research team announced that they were working with the Civic Museum in Bologna along with two other institutions, one in Rome and the other in Parma. Their project involved the use of video cameras placed around the galleries of all three institutions to monitor visitors. However, rather than simply tracking visitors to see which works of art they were most drawn to, the technology made use of AI algorithms to judge the emotions of members of the public as they interacted with the art on show.

Therefore, the AI project is much more than working out which parts of a museum’s collection are the most popular but deciding the degree to which people engage with them. By gauging reactions to individual items on display, such as works of art, the researchers hope to build an AI system that has some degree of empathy – or, at least, understanding – of the way people feel about what is before them. What’s more, this process is supposed to go on without the visitors realising that their emotions are being monitored even if they are aware of the presence of cameras.

Facial Expressions and Value Judgements

Without filling in questionnaires about their emotional reactions to pieces, the technology takes over and assigns certain value judgements, or scores, to what it detects. AI systems can already determine five distinct facial expressions and this is what the research team is using to work out how people feel. The algorithms used are able to work out when happiness or sadness is detected in a facial response as well as whether people feel surprised or angry. More subtly, however, the technology is also supposed to be able to work out when someone has had a neutral reaction to what they’ve seen. The key here, of course, is to determine whether someone is genuinely neutral about something emotionally or has simply not engaged with it at all.

Furthermore, the AI systems being trialled in the three Italian cities are not limited to judging people’s emotional responses to museum collections. They are also being used to make certain demographic judgements. For example, the technology now being operated is purported to be able to determine a museum visitor’s perceived gender as well as how old they are likely to be. Furthermore, the AI system is also able to capture information about people’s eye movements. According to the researchers, this will help the system to work out how much an exhibit is genuinely prompting interest as opposed to all of the other things you find in galleries, such as way-finding information and so on.

Computers: the Curators of the Future?

Although the experiments in AI are at the stage of gathering data and improving algorithms, the Italian technology specialists hope that the information gathered will be used to redesign museum layouts for optimal visitor experiences. This could, of course, include which exhibits are put on show and which ones are removed or placed in less prestigious settings. It seems, therefore, that what was once the job of curators could be automated. Certainly, this would look more likely if automatically curated galleries started to perform better than other institutions in terms of visitor numbers.

Silvia Battistini, a curator at the Civic Museum in Bologna, did not appear to be unduly concerned about the project or her future when she spoke to the press, however. Instead, she expressed surprise at what she deemed to be an opportunity to learn more. She said that as a curator she was not used to be able to get her hands on such useful information about how the public interact with the artefacts and artworks in her museum and that it could be a powerful tool in determining future exhibitions.

How this might work exactly as well as the privacy concerns some have raised over such close monitoring of visitors remains an open question, however. Some artists have already said that they don’t approve of the approach now being experimented with. In the main, they have complained that AI used in this way will become a race for more, not better, interactions, thereby devaluing art and museum exhibitions to the level of social media.