How to take AI far beyond gaming

AI.shutterstock_424934146


Venture capital investments in artificial intelligence (AI), one of the biggest technological trends these days, are booming. The year 2016 saw almost ten times more funds invested in the space than 2012 did. Virtual and augmented reality (AR/VR), sometimes associated with the creative side of AI, also became a hot topic: More than $1.8 billion was invested in AR/VR, compared with just $86 million in 2012.

Applying AI to solve tasks normally considered creative has historical precedent. For example, procedural generation has been used to draw textures, produce 3D models, and automatically generate large amounts of content in video games since the 1980s. Moreover, gaming has contributed significantly to the whole AI landscape. Revenues generated by game developers allowed chip makers like Nvidia to improve hardware, and research on classic Atari games became one of the cornerstones of DeepMind’s breakthrough in reinforcement learning. However, a question arises — is gaming the only industry where AI and creativity merge?

Luckily, cultural and creative industries (CCIs) are much more than gaming. CCIs, defined as industries whose workforce possesses creative skills required “to yield either novel, or significantly enhanced products whose final form is not fully specified in advance” generated £84.1 billion of gross value added in 2014 in the U.K. alone. Nine sectors, including advertising and marketing, architecture, IT, software, and computer services, grew 1.4 times faster than the U.K. economy between 1997 and 2014.

Tech giants also show how to apply AI muscles to a variety of sectors within the creative field. Some like Amazon, Etsy, Netflix, and Spotify use AI heavily to improve the discoverability of diverse content. Meanwhile, others like Microsoft, IBM, Alphabet, DeepMind, and Adobe push the envelope further. Music generation, graphic design, and even culinary arts have seen the involvement of AI.

How culture can benefit from AI

It’s no surprise that tech leaders are eager to combine AI and CCIs. Historical patterns suggest that cultural and creative industries are beneficiaries of adopting tech. For example, the previous industrial revolution, alongside electrification and factories, brought with it a plethora of new cultural phenomena like cinemas, amusement parks, and music halls. Similarly, the AI revolution may also enable new modes of cultural production, dissemination, and enjoyment, making on-demand, personalized artworks commonplace or enabling us to step into our own past and future through procedurally generated experiences.

The general public favors applications of AI that resonate with culture and creativity. For example, even though public interest in AI is gradually increasing, it spikes when AI concerns creative, cultural, or art-related issues. A sci-fi psychological thriller about AI, an app that edits photos, or an AI that wins an ancient game all fascinate people.

From a technical and scientific standpoint, creativity is “one of these last frontiers of AI,” says Hod Lipson, professor of engineering at Columbia University. Advancements in machine learning, computer vision, natural language processing, and other domains provide us with technological tools that can be used creatively and for artistic effect.

Convolutional neural nets power Google’s DeepDream algorithm, which enhances patterns in images, finding dogs and faces where there are none and creating psychedelic imagery. Researchers from Tübingen have used similar technology to develop a style transfer algorithm, turning images into the style of Monet or Picasso and inspiring consumer applications like Prisma.

Meanwhile, many in the AI research community have shifted their attention to generative adversarial networks (GANs). In their effort to improve and stabilize this technology, researchers created algorithms that occasionally produce surrealist-like imagery, which artists have started to adopt in their work and consumer apps are sure to follow soon. As for text and music, open source implementations of recurrent neural networks have facilitated the creation of rap music, cooking recipes, short films, and folk songs.

Therefore, the scale of the opportunity, the historical evidence, the natural interest of the public, and the technological advances make creative AI an interesting sector for entrepreneurs to approach. Let’s research some existing applications of AI in cultural and creative industries to map the terrain.

Current creative applications of AI

By analyzing 94 companies and corporate or academic initiatives, we can find at least four distinctive applications of AI to creative industries: content search and discovery, personalization, interaction, and creative process augmentation/automation.

Dealing with creative content is a difficult task, for professionals such as social media marketing managers and for consumers. For example, it is not humanly possible to watch all 400 videos uploaded every minute on YouTube. AI for the discovery of books, films, or artworks is actively used by large companies like Amazon or Netflix, as well as startups like Artfinder.

Products for creative professionals go one step further and involve humans into the traditionally black-box process of AI-powered discovery. Startups like FindTheRipple and Oz Content make it easier to navigate the ocean of content based on parameters such as sentiment, theme, or virality.

Adding a degree of personalization is another application of AI in cultural and creative industries. As direct products of human efforts, products of CCIs are rarely customizable. No artist can create personalized works at scale and at reasonable cost. This is where machines may help. Whether by creating augmented music that captures the state of the listener or by making an original mashup out of several music tracks, a flavor of individuality can be added to products of human creativity.

AI may help CCIs to become not only more personal, but also more interactive, empowering both consumers and artists. For example, Picasso Labs and Echobox make a feedback loop for creators, giving them information on how the public is reacting to attributes of their headlines or videos. Other AI tools create additional interfaces for human-computer interaction. Tagging any element of a video or turning any surface into a music instrument are made possible by Wirewax and Mogees respectively.

AI enthusiasts go further than fine-tuning products of human artists; they use technology to automate and augment the creative process. Generative AI applied to music, photos, video, and text produces original artifacts. Jukedeck and Mubert deliver original, royalty-free tunes, while Prisma and Soloshot make photographic miracles. Tools such as Wordsmith from Automated Insights construct written narratives from data, and experiments employ AI to compose Chinese poetry, write film scripts, and generate plots for musicals.

Open opportunities for artistic AI

Some niches may be still be terra incognita, where creative AI penetration is slowed down by some technical or business-related issues. For example, developing applications that require interactions between humans and AI is more difficult than developing AI itself. David A. Mindell, a professor of aeronautics and astronautics, and Dibner Professor of the History of Engineering and Manufacturing at the Massachusetts Institute of Technology, notes: “The most advanced (and difficult) technologies are not those that stand apart from people, but those that are most deeply embedded in, and responsive to, human and social networks.”

In some cases, applications of creative AI remain unfulfilled because of finances, not technology. For example, publicly funded or nonprofit institutions like museums and libraries may lack the resources required to innovate.  Moreover, despite their indisputable creative and cultural essence, some CCIs execute a limited number of actual creative tasks as measured by creative intensity (the proportion of creative jobs within an industry). When the creative intensity is low, applications of creative AI may be limited.

Overall, looking at multiple examples of applying AI to cultural and creative industries, it is justifiable to say that there are many opportunities outside the gaming industry. Entrepreneurs thinking about launching a creative AI business or integrating AI into an existing creative venture may spot an opportunity by looking at it from four perspectives: accessibility and discoverability, personalization, interactivity, and the creative process itself. Chances are high that the creative industry chosen is prone to disruption in one or more ways. And the rewards? Apart from financial returns, your startup may well contribute to humanity’s next major cultural breakthrough.

Peter Zhegin is an associate at Flint Capital, a venture capital fund with an exposure to cognitive tech. 

Luba Elliott is a creative producer, artist, and researcher exploring the role of artificial intelligence in the creative industries.

2017-03-14T20:01:46+00:00

Leave A Comment