A new creation engine? Bookmark has been added
A new creation engine?
Using AI to create original works
Wondering what’s next for artificial intelligence (AI)? Look for companies’ AI’s focus to expand from “optimization” to “creation”.
June 26, 2019
Source: Deloitte “State of AI in the Enterprise, 2nd Edition” survey, October 2018.
In a recent global analysis based on Deloitte’s State of AI in the Enterprise, 2nd Edition survey, AI early adopters were asked to identify the primary benefit of implementing AI in their organization.1 The top externally focused benefit was “enhancing existing products and services” (with 43 percent ranking it in their top three benefits). The top internally focused benefit was “optimizing internal business operations” (with 41 percent ranking it in their top three). In the current wave of AI adoption, organizations seem more focused on improving what they have, as opposed to creating something new.
However, there are signals that AI implementations may be expanding beyond enhancement and optimization: Individuals and companies are increasingly looking to use AI in the creative process. This includes creating new content with minimal human input.
Over the past few years, a growing number of experiments in this space have garnered public curiosity and media attention. While some could be considered publicity stunts, others have been dedicated attempts to advance AI’s capabilities.
For example, in 2016, IBM Watson created the first film trailer to be generated by artificial intelligence.2 It was trained on 100 different horror movie trailers and then, through that filter, it analyzed the movie for which it wanted to make a trailer. In the art world, some organizations, including The Art and Artificial Intelligence Laboratory at Rutgers, are exploring the use of AI algorithms to generate new art.3 AI is even being used to create music. OpenAI’s MuseNet is a new AI-powered tool that can generate songs with multiple instruments and in various styles.4 This tool joins a growing number of music-generating services, startups, and artists leveraging AI in their work.5 AI has also been used in journalism, poetry, sports, video game development, culinary arts, and other creative pursuits.
These developments are certainly exciting, but how can AI content creation move from behind the scenes to the main stage? There are likely several challenges to address.
First, for something to be truly creative, it should be both novel and useful.
Second, the appreciation of content is subjective and relies on the preferences, emotions, and experiences of the consumer. Organizations that use AI to generate content should make sure the human element isn’t lost, and that their content makes an emotional connection.
Finally, there are intellectual property and content ownership questions pertaining to AI-generated content.6 Current regulations vary from country to country, and these frameworks will be tested in the future. The United States Patent and Trademark Office has begun to explore this issue in depth over the past few years.7
Media and entertainment companies (and other content-intensive organizations) that are currently using AI to enhance their existing operations should consider experimenting with using AI for content creation—while being mindful of the potential challenges.
This charticle authored by David Jarvis.
1 “Future in the balance? How countries are pursuing an AI advantage: Insights from Deloitte’s State of AI in the Enterprise, 2nd Edition survey,” Deloitte Insights, May 1, 2019.
2 “IBM Watson creates the first AI-made film trailer—and it's incredibly creepy,” Wired UK, September 2, 2016.
3 “The AI-Art Gold Rush Is Here,” The Atlantic, March 6, 2019; The Art and Artificial Intelligence Laboratory at Rutgers homepage.
4 “OpenAI’s MuseNet generates AI music at the push of a button,” The Verge, April 26, 2019.
5 “12 songs created by AI,” Google Arts & Culture.
6 “Legal issues around IP for AI: Who owns the copyright on content created by machines?”, Techworld, January 26, 2018.
7 “Remarks by Director Iancu at the ‘Artificial Intelligence: Intellectual Property Considerations’ event,” United States Patent and Trademark Office, January 31, 2019.