Few people realize that artificial intelligence has already entered the museum world. It is one of the underlying technologies powering the many trends and innovations in the field, such as big data and robotics. AI-enabled virtual humans have been touring visitors around museums for years, such as in the Museum of Science in Boston. The Dallas Museum of Art, Cooper Hewitt Museum and many other museums have started to analyze big data, more or less with the help of AI, to understand museum collections. Some predict that artificial intelligence will reshape businesses outside of IT and become a competitive asset in 2016. The Boston start-up Cuseum, which offers mobile technology solutions to museums, recently published an explorative post on museums and AI. However other than these examples, this cutting edge technology has not been given enough attention by museums. How can museum experts envision the future that may not be too far away with smart machines? More specifically, how can artificial intelligence enhance museum visitors’ experience? Here are 4 possible ways in which AI can transform museums through accessibility and visitor experience improvements:
1. Captioned images for the vision-impaired
How can artificial intelligence improve accessibility for the vision-impaired? The Metropolitan Museum of Art in New York and many other museums have already made “touching the art” possible for the vision-impaired with the help of 3-D printing. MoMA also prepares lectures with visual descriptions of artwork and provides specially trained lecturers who lead touch tours. However, all the artworks have to be selected and pre-prepared by the museum. Currently, the vision-impaired have a tool called screen reader that can read out the content on the screen. For example, if the label reads “Venice, from the Porch of Madonna della Salute; Artist: Joseph Mallord William Turner (British, London 1775–1851 London); Date: ca.1835,” the screen reader will read out each word exactly. With the object recognition function powered by AI, it will read out not only the label content, but also caption as “The image contains black boats, blue sky, white clouds and buildings…” This technology has already been developed by Facebook to make vision-impaired users feel more engaged with posts. Museum audiences might also benefit from this idea; they will be able to explore the many other artworks that no longer need to be pre-prepared by museum staff.
2. Collection analysis
Machine learning has been used in analyzing paintings for some time. Babak Saleh and his colleagues at Rutgers University in New Jersey have classified paintings according to color, line and objects shown in the piece, and found connections between them. Bengler applies machine learning and deep neural networks to the Norwegian National Museum’s collections to lay out maps of paintings based on their perceived similarity. These projects are surprising and inspiring and evoke new understanding on the artworks. However, while many of the works are novel and explorative, museum experts are skeptical about the methodology. Is it a genius path to find patterns in art or does it take the joy out of interpretation? How to interpret and apply these works? More discussion from museum experts are expected on this thriving research topic.
3. Online store and customer service
Almost every museum has an online store, but at the end of the day, not every store satisfies customers. Online retail giants like Amazon take advantage of artificial intelligence and business analytics to gain insight from the behaviors of consumers. Imagine a customized museum online shopping experience. An AI interface could carry on sophisticated conversations with the customer and lead him to the very item he wants, all the way till the transaction is completed. This system could also follow up with the customer after the sale, and store customer information for the development department. Since “millennials don’t want more customer service—they want different customer service,” why not apply systems like Siri and Cortana to help customize future online museum retail experiences?
4. Translation and communication
Every year, more than 6 million foreign visitors visit the Louvre. A distinct language barrier manifests when there are not enough volunteers and audio devices serving this host of special language needs. Microsoft recently added a new feature in Skype to translate between languages (almost) immediately during the call. Skype Translator will improve with more uses since it is powered by machine learning. Similarly, we can place AI kiosks or humanoid robots in museums as personal language assistants, as well as tour guides. For example, Berenson, the robot from the Musée du quai Branly in Paris, named after the late critic Bernard Berenson, was programmed to demonstrate its feelings towards the art through facial expressions. We can then expect to see robots like Mr. Berenson ”talk” about their feelings in different languages one day.
There are other applications as well. During the 2016 Museums and the Web conference, the museum ticketing company Area360’s CEO Chris Smith demonstrated how to book a Laker’s ticket through Amazon Echo, a far-field voice control gadget and powered by voice-controlled AI named Alexa. He expressed the desire to bring this experience to the museum world. Imagine next time you want to visit a museum, just ask: Echo, I want to visit the Guggenheim Museum on Tuesday, plan the trip, avoid peak times, buy two adult tickets and borrow a baby stroller—then you are all set.
Though artificial intelligence systems have the ability to improve visitor experience in many ways, it doesn’t mean that museums should replace or reduce traditional labor. Rather, museums that adopt artificial intelligence systems might trim operational redundancies and make the organization perform more efficiently. Indeed, humans will still be the ones to attend board meetings, make strategic plans, solicit artworks, curate exhibitions, and experience the art.
Artificial intelligence is more than cute or funny robots; it is also a new equipment for us to learn and use data, such as languages, sounds and voices and images.
Images licensed under creative commons. By Audrey Defretin - Own work, CC BY-NC-SA 2.0