Carnegie Mellon University, Heinz College of Information Systems and Public Policy
By Jennifer Monahan
Is art inherently better because it is created by a human?
When asked that question, Brett Ashley Crawford, Ph.D., faculty chair of the Master of Arts and Entertainment Management programs, joint efforts between Heinz College of Information Systems and Public Policy and the College of Fine Arts at Carnegie Mellon University, answered with a scenario and some questions of her own.
“Consider a wonderful writer, like [travel writer] Bill Bryson and how his position and journey (born in Iowa to a father who was a journalist), and circumstances, afforded him the opportunity to backpack through Europe with a friend. Going to all these places – which is very inspirational – fuels you as an artist,” Crawford said. “If I’m somebody who has a physical disability that prevents travel, or if my life journey has different financial or systemic challenges that prevent me from following that same path as Bryson, maybe using technology can help me get to something different, with a different form of inspiration. Is that better or worse, for the artist? The artist has been inspired to create something with AI. Is being inspired by AI really any different than the way Bill Bryson was inspired by the ability to do what he did?”
The complexity of the response is indicative of the impact artificial intelligence (AI) is having in the world of art and entertainment. At odds in the scenario above are two things that both ring true, but don’t fit well together:
the presumption that human creativity is predicated on authentic, lived experiences;
the idea that AI can assist human creativity in a way that makes creation more accessible for more people, no less authentic nor less valuable just because it’s digital.
Daniel Green, Ph.D., director of the Master of Entertainment Industry Management Program, raised similar questions about the nature of creativity.
Though the quality of plays, scripts, or short stories produced by generative AI may not yet be equal to that of a seasoned writer, it’s good enough that protection from competition by future generative AI tools is one of the issues being debated as part of the current writers’ and actors’ strikes in the U.S. The threat of a technological tool that can take in a popular book and instantaneously churn out a passable movie screenplay based on that story has writers worried.
“For years, critics have been wary of entertainment – dramas or romantic comedies, for example – that are formulaic. They’ll say the work is derivative,” explained Green.
Classic literary plotlines are fair game and have influenced many modern movie successes. The Lion King is loosely based on Shakespeare’s Hamlet. Clueless follows Jane Austen’s Emma. 10 Things I Hate About You was inspired by The Taming of the Shrew (more Shakespeare).
“Is generative AI really doing anything different?” Green asked. The examples provided by Crawford and Green highlight a central dilemma caused by AI in the world of art and entertainment: that seemingly conflicting elements are both true. AI presents paradox after paradox, and the use of AI in art raises tangled questions with no easy answers.
Authorship and Transparency
Chief among these concerns is the question of authorship. When an artist or songwriter uses AI to create something, who owns the content?
In April, the AI-generated song “Heart on My Sleeve” created a huge buzz. The song featured a fake Drake and The Weeknd, stealing their sound and style. While those musicians didn’t write the song, its popularity relies on their talent and fame. So who gets credit and compensation? Drake and The Weeknd? The person who created the prompt and put the song together? The company who created the AI tool? The computer engineers who wrote the algorithms for that model? The artists who created the content that the AI was trained on? Depending on how one defines authorship, the answer to all of those questions could be “yes.”
The same issue exists in visual arts.
“Artists playing with AI have existed on the fringe for many years,” Crawford said. “With digital artwork, I think there has been this assumption that we know technology has been involved and we just let it go, because it was truly more like having a better paintbrush. But AI has gotten better, and artists today are using it differently.” - BRETT ASHLEY CRAWFORD, PH.D.
The quality of current AI is more like collaborating with a co-artist than having a better paintbrush. With that development, Crawford said, she expects to see more transparency about the authorship of a work of art. Competitions will have guidelines clarifying whether and how much assistance from AI-tools is acceptable. Finished pieces hanging in museums or galleries will include descriptions that detail the creation process and how AI was used.
Green envisions a similar future in the entertainment industry. “Writers won’t be happy, but I could imagine there will be an acknowledgement that a story was written in combination with AI. So, for example, the credits would list the AI program used along with the names of the two writers who rewrote the AI-generated script,” Green said.
While people might be more familiar with high-profile stories like “Heart on My Sleeve” or Jason Allen’s first-place prize in the 2022 Colorado State Fair art competition for the AI-generated “Theatre D’Opera Spatial,” less public creative and administrative roles across the arts and entertainment sector are equally likely to be affected by AI.
Green referenced the AI-generated opening credits for Marvel’s Secret Invasion series. The use of AI in this case created some cool visual effects. It also took work away from the animators and graphic designers who would typically be employed for such a job.
Crawford, who is also the executive director of CMU’s Arts Management and Technology Laboratory, said the increasing use of The Volume technology to create digital scenery – think Star Wars: The Mandalorian – is already changing the job market for designers and location-based filming.
The recent “Joan is Awful” episode of Black Mirror hilariously – and disconcertingly – highlights the potential complications of what happens when actors license their digital image and likeness for use in new shows. Green speculated that audiences might someday see new films starring James Dean or Marlon Brando, though such conjectures prompt more questions about who is compensated when the featured digital actor is deceased.
While it is easy to focus on the potentially negative implications of AI, both Green and Crawford are balanced in spotlighting its merits.
Increased Access, Engagement, and Efficiency
For content creators, AI tools have mitigated some barriers to producing music.
“Musicians used to have to go into a studio. They used to have to know someone, or get a big break, in order to get studio time and resources. That’s difficult and expensive. Once the track was made, they had to figure out how to distribute it,” Green said. “Now, if you’re talented enough, you could record all those pieces onto a machine, use AI to mix it, drop in your vocals, and then you could distribute it an hour later on Spotify.”
For people with visual, auditory, or mobility challenges, AI can change the landscape – literally.
AI tools are already being used to enhance access and engagement. “Those same large language models that power ChatGPT can also be used to provide surtitles in operas for hearing-impaired patrons,” Crawford explained. “There’s new software that’s not only generating the text, but placing the words in the user’s line of vision right above the person who is speaking.”
Museums are doing inventive work with technology to improve the patron experience. Have you ever wanted to pick up a piece of art in a museum? Strictly verboten, of course, but a sophisticated 3-D view offers a similar experience. Innovations at the Cleveland Museum of Art allow anyone, anywhere, to zoom in and out on selected objects, flip them upside down, spin them around, and see them from every angle. Images are accompanied by a “Did you know?” display featuring fun facts, plus brief, interesting descriptions that can help the uninitiated learn more about the world of art. During Covid, the Tate Britain launched its “After Dark” website; the museum had four remote-controlled robots, and viewers could take turns controlling robots’ movements as they roamed the museum at night.
teamLab, an international art collective, explores the confluence of art, science, technology, and the natural world. Its exhibits often include immersive experiences allowing visitors to engage with the art and even create elements of it.
From an administrative as well as a creative standpoint, AI has much to offer arts managers. “Arts organizations have been using AI for a while to assist with provenance,” Crawford said, and law-enforcement agencies like the FBI can use image-recognition software to help identify stolen artwork.
In the non-profit world of museum management, AI tools can be a godsend. Arts organizations are often under-resourced and consequently understaffed. They may not have a marketing team with the bandwidth to create strategic campaigns. Customer relationship management (CRM) software can tailor marketing campaigns to the specific profiles of individual patrons, while chatbots such as ChatGPT or Bard can generate social media posts.
“Efficient and effective communication between the organizations and their audiences is going to become better,” Crawford said. With individualized marketing, however, come accompanying privacy concerns. Crawford cited examples of geofencing and interactive outdoor advertising as normalized marketing opportunities but potential pain points for an individual’s privacy. “The tensions are around personal data – what big tech companies are allowed to do with my data, and how they are allowed to do it,” Crawford explained.
Embedding Ethics Into Everything
As educators charged with training the next generation of arts and entertainment industry leaders, Crawford and Green grapple regularly with how to teach their students to negotiate the ethics of a world where the technology – and the impact of that technology – is changing so fast.
MEIM graduates work across the entertainment sector in streaming, gaming, filmmaking, music, and business. MAM graduates lead arts organizations, museums, theaters, and galleries. They are the current and future leaders of an industry that is being fundamentally changed by technology.
“The students are going to be using this technology with or without us,” Green said. “We have to teach ethics in the classrooms. We have to give them frameworks for how to navigate this reality.”
Faculty – even faculty at a university known for being at the forefront of new tech developments – cannot always anticipate what the upstream and downstream effects of generative AI will be, much less whatever technology comes after that. And after that. Instead of teaching students to respond to a particular situation, faculty teach them how to make decisions amid uncertainty and how to anticipate and consider the impact of those decisions for various audiences.
“Part of what I try to develop is the habit of ethics,” Crawford said. “As an educator, my approach is that ethics is simply part of every question we talk about. Arts organizations that are trying to be true partners in an equitable and inclusive world have to have high ethical standards and to recognize where the ethical choices are.”
“We have to be forward thinking with any policy around technology,” Green said. “Artists will increasingly use AI, and things are going to continue to change quickly.” To that end, the MEIM program is launching a fall class focused on AI in entertainment. This elective class will include an experiential learning component that allows students to create something using AI.
“We have to be nimble enough to pivot our curriculum to what’s happening in the entertainment industry,” Green explained. “Media and entertainment disruption will only be exacerbated by the use of AI.” - DANIEL GREEN, PH.D.
“Boundaries around AI need to be created sooner rather than later,” Crawford said. “Because AI has some powers that are currently unchecked and perhaps not understood, our issues with privacy and security need to be addressed first.”
Policymakers also need to consider appropriate compensation as well as ownership and authorship transparency with respect to generative AI. Copyright law needs to address the role of collaborative AI – that is, when a human artist uses AI to create art – and how ownership in such a case will be codified into law.
Crawford noted that international copyright laws will likely be established first and that the U.S. should follow suit. As an example, both Japan and Korea have a robust art and technology interface as well as interesting approaches to ownership of AI-generated art. The European Union is moving forward with an AI Act that includes provisions for copyrighted material.
While no one can predict the future, Crawford believes that the real game-changer is not AI, but quantum computing.
“Quantum computing is still in the developmental stage,” Crawford explained. Estimates are that it might enter the marketplace anywhere from 2025 to 2035. And that, Crawford said, will change the entire ecosystem.
In the meantime, AI offers plenty to grapple with.
Like every tool, AI has the potential to help and harm. In the context of art and entertainment, AI creates new opportunities and provides broader accessibility. It also requires thoughtful consideration about how it can be deployed in ways that assist human creativity rather than replace it, and how it can be used ethically, to contribute to the common good. Those thorny issues won’t get solved overnight. AI will most certainly be part of the ongoing story in art and
entertainment; learning to reconcile its paradoxes is the new creative challenge.