If you were to ask people how AI could change their lives, they may immediately think of self-driving cars and chatbots. In a business context, increased efficiencies and advanced data analytics would be among the likely responses.
But AI is also changing the arts, enriching people’s daily experiences, preserving culture and making art more accessible to those unable to visit a gallery or historic site for themselves.
In July 2019, Microsoft announced a new and fourth pillar to its AI for Good portfolio, the $125 million, five-year commitment to use artificial intelligence to tackle some of society’s biggest challenges. This new pillar will focus on AI for Cultural Heritage, and use AI to work with non–profits, universities, and governments to help preserve the languages we speak, the places we live, the artifacts we treasure and celebrate the people who have made an impact.
The program will build upon previous efforts including those in New York, with The Metropolitan Museum of Art and MIT; in Paris with the Musée des Plans-Reliefs; and in southwestern Mexico, where Microsoft is engaged as part of ongoing efforts to preserve languages.
AI is also a creative force able to compose music, write novels and paint pictures. Here are seven examples of how AI is enriching our cultural lives.
The great buildings and historical sites of the world may attract millions of tourists a year, but many more people have only seen pictures. That is beginning to change.
Microsoft AI is being used to help preserve records of historic sites and bring people closer to some of the wonders of the world. Teams from the French company Iconem have used cameras and drones to create 3-D digital models of landmarks from Cambodia to Syria.
In Paris at the Musée des Plans-Reliefs, Microsoft partnered with Iconem and HoloForge Interactive to create an immersive experience using mixed reality and AI that pays homage to the French cultural icon Mont-Saint-Michel, off the coast of Normandy.
Visitors can interact with the exhibits and discover information and stories about the site.
The Metropolitan Museum of Art launched its Open Access initiative in 2017, making all images and data relating to public-domain artworks in its vast collection available to everyone online. The Met recently collaborated with Microsoft and the Massachusetts Institute of Technology to help take this initiative to the next level, using artificial intelligence to explore new ways for global audiences to discover, learn and create with one of the world’s foremost art collections.
There are more than 7,000 languages in the world, a third of which have fewer than 1,000 people who continue to speak them. In southwestern Mexico, Microsoft is engaged as one of the community partners in efforts to preserve languages spoken in the region, specifically Yucatec Maya and Queretaro Otomi. By using AI, Microsoft has helped to protect endangered languages.
[Subscribe to Microsoft on the Issues for more on the topics that matter most.]
Enigmatic expressions like those worn by the “Mona Lisa” or seen on the faces of countless statues of Buddha, invite the viewer to speculate on what the subjects was thinking or feeling.
But researchers in Japan are revolutionizing the way we think about this phenomenon, using facial recognition software. They used Azure Cognitive Services Face API to analyze 200 statues of Buddha, including the mysterious expressions of the Ashura Buddha at the Kofukuji Temple in Nara.
Traditionally, Buddhist statues would have shown faces devoid of emotion. But in their creation, the Kofukuji statues’ faces may have been influenced by their sculptors’ moods and may carry traces of detectable emotion, which the project sought to investigate. The aim of the project was “to provide people with a means for reaffirming the beauty of Buddhism,” according to Professor Syun’ichi Sekine.
Beyond Microsoft’s efforts, these are just some of the ways AI is already changing the arts:
In 2019, OpenAI announced that it had created a language algorithm that could write text that was indistinguishable from that written by a human. The GPT-2 program has not been released as a fully trained version, as its makers claim they are concerned about the potential “malicious applications of the technology.”
Whether that concern is justified, AI is already writing both news and fiction. In 2014, the Los Angeles Times reported on an earthquake that had just hit the city, with an article automatically generated by its Quakebot algorithm. And Guardian Australia has run an experiment in publishing an article written by a program called ReporterMate. Such developments are intended to produce straightforward news items with as little human intervention as possible, leaving editorial staff and reporters free to focus their efforts on more complex or nuanced activities, such as investigations or opinion pieces.
AI has also been credited with writing its first novel, “1 the Road,” an account of a road trip written by a computer hooked up to a GPS, microphone and camera.
Painting by numbers
In October 2018, the sale of the painting “Portrait of Edmond de Belamy” for $432,500 surprised the art world. The “artist” was an algorithm used by the Paris-based collective Obvious. Members of the collective fed thousands of images into a computer, which then used what it had learned to create an original image.
The sale sparked debates about what constituted art and whether human artists would eventually be replaced by machines. But the people behind Obvious are far from the only ones using AI to create works of art. The HG Contemporary Gallery in New York has hosted an exhibition called “Faceless Portraits Transcending Time,” featuring prints produced by an AI program named AICAN.
Composing melodies and lyrics
AI has been used in music for decades. In 1956, Lejaren Hiller used a computer to help compose the “Illiac Suite for String Quartet.” And the influential producer Brian Eno help pioneer a genre called generative music.
Today, AI is being used to write so-called functional music for commercial clients like the video game industry, with tempo and mood configured to keep up with changes in ongoing gameplay.
Another impact AI is having on music is in the use of algorithms that create playlists on streaming services. Not only are they choosing what millions of subscribers listen to, but they are also beginning to introduce AI-written music into those playlists.
For more on AI, visit AI Empowering Innovation. And follow @MSFTIssues on Twitter.