How Artificial Intelligence and Open Data can re-imagine our cultural future

We are currently in the midst of an unprecedented race to master innovations in Artificial Intelligence. Over the past year, the star of the show has been Generative Artificial Intelligence (GenAI), i.e., that which is capable of generating original and creative content such as images, text or music. But advances continue to come and go, and lately news is beginning to arrive suggesting that the utopia of Artificial General Intelligence (AGI) may not be as far away as we thought. We are talking about machines capable of understanding, learning and performing intellectual tasks with results similar to those of the human brain. 

Whether this is true or simply a very optimistic prediction, a consequence of the amazing advances achieved in a very short space of time, what is certain is that Artificial Intelligence already seems capable of revolutionizing practically all facets of our society based on the ever-increasing amount of data used to train it.  

And the fact is that if, as Andrew Ng argued back in 2017, artificial intelligence is the new electricity, open data would be the fuel that powers its engine, at least in a good number of applications whose main and most valuable source is public information that is accessible for reuse. In this article we will review a field in which we are likely to see great advances in the coming years thanks to the combination of artificial intelligence and open data: artistic creation. 

Generative Creation Based on Open Cultural Data  

The ability of artificial intelligence to generate new content could lead us to a new revolution in artistic creation, driven by access to open cultural data and a new generation of artists capable of harnessing these advances to create new forms of painting, music or literature, transcending cultural and temporal barriers.  

Music  

The world of music, with its diversity of styles and traditions, represents a field full of possibilities for the application of generative artificial intelligence. Open datasets in this field include recordings of folk, classical, modern and experimental music from all over the world and from all eras, digitized scores, and even information on documented music theories. From the arch-renowned MusicBrainz, the open music encyclopedia, to datasets opened by streaming industry dominators such as Spotify or projects such as Open Music Europe, these are some examples of resources that are at the basis of progress in this area. From the analysis of all this data, artificial intelligence models can identify unique patterns and styles from different cultures and eras, fusing them to create unpublished musical compositions with tools and models such as OpenAI's MuseNet or Google's Music LM. 

Literature and painting  

In the realm of literature, Artificial Intelligence also has the potential to make not only the creation of content on the Internet more productive, but to produce more elaborate and complex forms of storytelling. Access to digital libraries that house literary works from antiquity to the present day will make it possible to explore and experiment with literary styles, themes and storytelling archetypes from diverse cultures throughout history, in order to create new works in collaboration with human creativity itself. It will even be possible to generate literature of a more personalized nature to the tastes of more minority groups of readers. The availability of open data such as the Guttemberg Project with more than 70,000 books or the open digital catalogs of museums and institutions that have published manuscripts, newspapers and other written resources produced by mankind, are a valuable resource to feed the learning of artificial intelligence. 

The resources of the Digital Public Library of America1 (DPLA) in the United States or Europeana in the European Union are just a few examples. These catalogs not only include written text, but also vast collections of visual works of art, digitized from the collections of museums and institutions, which in many cases cannot even be admired because the organizations that preserve them do not have enough space to exhibit them to the public. Artificial intelligence algorithms, by analyzing these works, discover patterns and learn about artistic techniques, styles and themes from different cultures and historical periods. This makes it possible for tools such as DALL-E2 or Midjourney to create visual works from simple text instructions with aesthetics of Renaissance painting, Impressionist painting or a mixture of both. 

However, these fascinating possibilities are accompanied by a still unresolved controversy about copyright that is being debated in academic, legal and juridical circles and that poses new challenges to the definition of authorship and intellectual property. On the one hand, there is the question of the ownership of rights over creations produced by artificial intelligence. On the other hand, there is the use of datasets containing copyrighted works that have been used in the training of models without the consent of the authors. On both issues there are numerous legal disputes around the world and requests for explicit removal of content from the main training datasets.  

In short, we are facing a field where the advance of artificial intelligence seems unstoppable, but we must be very aware not only of the opportunities, but also of the risks involved. 

Content prepared by Jose Luis Marín, Senior Consultant in Data, Strategy, Innovation & Digitalization. The contents and points of view reflected in this publication are the sole responsibility of its author.