This episode has been a long time coming. I have been trying to figure out how to tackle a discussion about A.I. and it’s influence on how sound editors will all be able to do our work in the future. So when the MPSE put out a press release about A.I. created work and it’s eligibility at their Golden Reel Awards, I saw it as a chance to broach the subject. David Barber, MPSE President, and Brent Findley are both on the A.I. committee with the guild and have spent countless hours thinking about how A.I. is going to effect sound editors.
Sponsor:
Sound Ideas is Introducing the newest instalment to The Hollywood Edge’s legendary sound effect collection: Premiere Edition – Volume 12! Build a rich audio-sphere to enhance all your productions, with this impressive assortment of sound genres: ranging from Machinery, to Nature, to Science Fiction. Access over 3000 sounds, available in four resolutions and with descriptive metadata in UCS formatting.
Head over to https://www.sound-ideas.com/Product/2461/Premiere-Edition-Vol-12-Sound-Effects-Library to learn more.
MPSE PRESS RELEASES ABOUT AI:
Here is a copy of the Motion PIcture Sound Editors’ original press release:
MPSE Bans Generative A.I. Sound from Golden Reel Award Eligibility
Los Angeles, Calif. – June 12, 2025 –
The Motion Picture Sound Editors (MPSE) has always promoted and will continue to value the human endeavor of artistic creation. Our Golden Reel Award honors people who utilize their skills and artistry to push the boundaries of storytelling through sound.
We support and prize technological advances that assist artists in their creations. However, standards for the legal and ethical use of Generative A.I. have yet to be established and are far from being accepted broadly. What we choose to promote as award-worthy points to how much we value the human endeavor of artistic creation.
Because of these concerns and our organization’s mission to uphold and celebrate the creative human endeavor of sound and music editorial, sound design, and Foley artistry, the MPSE will not be accepting projects for awards consideration where Generative A.I. is used to create elements for the final soundtrack. Golden Reel Award qualification rules are being updated and
will be published on our website soon.
As the industry evolves to adopt tools and techniques, current or yet to be developed, our board of directors and the A.I. Committee are poised to evolve our stance on this issue, ensuring we adhere to our mission. We remain open to what the technological future may bring in support of the humans working in the craft of sound editorial.
“It is an enormous question to ask: how much of our humanity are we willing to give away to technology, especially in the arts? The time to ponder that question, set up boundaries, and guide how A.I. is assimilated into our workflow and lives, was yesterday. The dam of A.I. has broken, and the waters are upon us. Choosing what we embrace as award-worthy filmmaking is
a way of diverting those waters while we grapple with this exponential change. As A.I. technology infiltrates and permeates our industry, rules and accepted practices for its use need to emerge that keep artists at the forefront.”
– David Barber, MPSE President
Below is the follow up article from MPSE, with more information dated June 23rd:
| Generative AI: Taming the Dragon A little over a week ago, the Motion Picture Sound Editors (MPSE) announced that we would not permit Golden Reel Awards consideration for projects that utilize Generative AI. We’re taking this opportunity to expound upon and further clarify our position. Our stance stems from the deeply unsettled legal and ethical landscape surrounding Generative AI within our industry. The MPSE’s Rules and AI committees have and continue to approach this topic with great solemnity and care. The current and future role of AI, particularly Generative AI and, imminently, Artificial General Intelligence (AGI), represents a profound challenge to the art and craft of sound, the film industry, and society at large. Our industry has largely reacted to this technology with apprehension, ambivalence, or worse, a laissez-faire attitude, conceding AI’s integration into our culture as a fait accompli to be dictated by the technology itself and those creating it. Our decision initiates a crucial pause. We are reclaiming some control and seeking to establish norms and standards that we deem essential for our awards consideration. GENERATIVE AI AS A TOOL We refuse to evaluate Generative AI models solely based on developer claims or their superficial convenience. Our assessment demands a deeper understanding of these models’ full potential, both beneficial and detrimental. Our fundamental principle for evaluating AI technology in the context of awards consideration is clear: Ethical AI – created and/or used with the intention to assist human artists UnEthical AI – created and/or used with the intention to replace human artists While this distinction may appear simplistic, it serves as a good springboard for analyzing the multitude of Generative AI products that have recently overwhelmed our craft. The line between these is extremely blurry but still visible. While we can still see it, we need to take action to define it before it gets defined for us. Through this lens, we can scrutinize the growing number of AI products targeting our craft and work to ensure AI functions as a tool for sound creatives. Machine learning models, LLMs, and Generative AI are advancing at an alarming rate. Most of us have interacted with at least one, if not all, of these types of AI. The interval between significant breakthroughs is shrinking, leaving us insufficient time to assess long-term effects or even to decide if we genuinely desire these integrations. This onslaught of advancements also cloaks the subversion of AI technology from tool to potential successor. While the number of AI products has exploded recently, we still have agency. We can and should control how we view and employ this technology. For the MPSE, that agency comes in the form of adopting rules not to award projects that utilize Generative AI outputs before we can lock down the parameters under which Generative AI or Generative AI-assisted work should be accepted. We must evolve our thinking on this together at a pace that we dictate, not one dictated to us by the proliferation of this technology. QUESTIONABLE ORIGINS: The Unethical Foundation The dirty little secret of many LLM, ML, NLP, and Generative AI models is that they consumed voluminous amounts of content, both fair use and copyrighted, by scraping the internet and other means to train their models. This unauthorized use of copyrighted works – books, art, paintings, music, scripts, etc. – has resulted in numerous pending lawsuits against tech companies. The most recent being Disney and NBCUniversal filing suit against Midjourney for copyright infringement. These lawsuits underscore not just legal questions but profound ethical ones as well. If the very foundation of these AI models is built on the unauthorized use of copyrighted works, it follows that any generated output constitutes “fruit of the poisoned tree.” And that material should not be allowed to be copyrighted, profited from, or, in our case, awarded. This is probably the most convoluted circumstance surrounding the current state of AI, as the illicitly obtained data is inextricably “baked into the cake” of these AI models. There is no simple way to “remove the egg” once the model is “fully cooked.” The die is cast. The original sin committed. Until these critical legal and ethical matters are resolved in courts of law, we cannot comfortably elevate the byproducts (i.e., generated outputs) of such models to award-worthy status. The MPSE is not saying that every Generative AI company has behaved unethically. We invite and encourage developers to participate in the discussion, providing transparent proof of their legal and ethical data acquisition. Furthermore, this collaboration can help us craft awards criteria that ensure their products serve as tools for our creatives, not as vehicles for the replacement of our craft or industry. Compounding this ethical conundrum is the effective transfer of knowledge and skill from experts and professionals to those with no experience in the field at all. Generative AI outputs are not genuine creations of the prompter; they are merely outputs for the prompter. This distinction is crucial. Our goal is to champion tools that serve to make people better artists and not those that merely facilitate the imitation of artistic creation. THE VALUE OF HUMAN EXPRESSION What we choose to promote as award-worthy points to how much we value the human endeavor of artistic creation. As Generative AI develops, it will be able to output “art” that is indistinguishable from human-created art. Music, video, photo-realistic stills, and other products will all replicate the original so well that we will be unable to tell the difference. While such quality might not be achievable today, the exponential growth of these models indicates it will be sooner than we think. This weighed heavily on our discussions: What, precisely, are we comfortable honoring? Consider this hypothetical: On one hand, legendary tenor saxophonist John Coltrane writes, rehearses, and records “Giant Steps” (a notoriously difficult set of chord changes to solo through), and that recording is subsequently submitted for an award. On the other hand, you have John Doe at his computer, typing in a prompt, “Jazz combo, Giant Steps, tenor sax solo, style of John Coltrane, 5 minutes,” and submitting the output for an award. Are we willing to say that these things are equal? Are we willing to put them side by side for award consideration? As an organization committed to the human endeavor of artistic creation through sound, we will not equate these. At this time, allowing unfettered submission of Generative AI outputs, in whole or in part, for award consideration would fundamentally undermine our values. We must work to implement clear standards and practices for the legal and ethical integration of Generative AI into our field. THE END GAME Our position is neither Luddite nor absolutist. AI is an undeniable, ubiquitous presence that will increasingly integrate into our professional and personal lives. We are not advocating against familiarizing or utilizing these new products as tools. What we are saying is that, with the jury being out (quite literally) on the legal and ethical implementation of these Generative AI products, the measure of creative credit is in question. Consequently, we will not wholesale accept AI-generated works for award consideration at this time. We are, however, researching and analyzing Generative AI tools and their place in our field as they relate to our awards. Our eyes are wide open, and we encourage upstanding companies that feel they ’ve been caught up in the net of unethical Generative AI to come to us and help us create these new norms. This invitation extends equally to audio professionals who are thoughtfully utilizing Generative AI as creative tools. We are starting a conversation that needed to happen yesterday. It is up to us to collectively decide whether to let Generative AI seep into our work and lives controllably or to allow it to wash over us. The choice and its consequences are significant. David Barber, MPSE President MPSE AI & Rules Committees |
LINKS: