Navigating the pros & cons of gen AI in creative workflows

by Kim Owens

On Thursday, May 9 at Blankspaces in Santa Monica a multi-faceted, multicultural group of creators, technologists, filmmakers, lawyers, students, and business folks gathered for "Gen AI in Creative Workflows: Balancing Innovation and Ethics,” to listen and learn from the two esteemed speakers.

Laura Randant Crenshaw is an entrepreneur, inventor, and the CEO of Mythulu, a company whose tools are designed to “challenge the stagnant pool of tropes and cliches that clutter your creativity” and Faith 'Aya' Umoh is a data scientist, media artist, and the founder of Creative Aya and ImmerseMe AI.

Odessa Reyno, Director of Women in Big Data SoCal chapter, began the conversation highlighting the premise not only of this evening’s topic but the mission for these conversations overall: bridging the gap between global resources and local need.

Reyno then introduced the evening’s moderator, Rachel Whaley, the data equity program lead for LA Tech4Good and partner presenter, who set the stage to discuss the three buckets of AI narratives. One: This is cool! Two: This is really scary. Three: AI is really interesting and opens up so many new possibilities, but what concerns should we be aware of and how do we best address these concerns and potential harms?

Is this a cool thing? AI as a creative tool and assistant

Speakers Aya Umoh and Laura Crenshaw with moderator Rachel Whaley

Some of the most common use cases for integrating AI into any workflow, creative or otherwise, are those tasks that may be repetitive, mundane, or time consuming and could be performed or automated by the right AI tool for a specific job. This could be transcribing a recording, as I have for the purpose of writing this recap, summarizing a large document, or generating key takeaways and next steps in just minutes following a Zoom meeting.

Aya shared how the use of AI tools like Prometheus in her creative process aids in developing 3D models and enhancing interactive experiences. She emphasized the importance of AI as a collaborator rather than a replacement of a human.

"You have an AI assistant to help you do it a little faster and help you to go through an extensive library. It’s for someone who's a novice at 3D art and who wouldn't necessarily know [how to do it]. It's the democratization of video games and 3D modeled videos. That kind of thing where you're going to start seeing a lot more deliverables,” Aya said.

Laura echoed this sentiment, explaining how her company Mythulu’s creativity cards, which help authors overcome their creative ruts, came to be. By using a forked version of Google Deep Dream, or what they call Art AI, they transformed art they had licensed into custom oil painting styles for the final card’s artwork. "It saved us 95 percent on our art costs out the door and allowed us to take a lot of art that was reconcilable and make it look cohesive,” she noted. “In all the years - I have sold many thousands of those card decks - I've had two people look at the art and be able to recognize any of the AI artifacts in it.”

This seems really scary: ethical considerations and copyright issues

Attendees mixing it up in conversation

Although this discussion took place before the ScarJo vs OpenAI privacy debacle, it was no surprise that a good portion of the discussion focused on the ethical implications of AI in creative fields, particularly concerning copyright and intellectual property.

Laura had a strong viewpoint after her deep dive into copyright law cases going back 180 years, including those that made it to the Supreme Court. Based on her extensive research, she believes the use of existing works to train frontier models such as OpenAI, MidJourney, and others often falls under fair use. That the essence of copyright exists only to protect blatant reproduction of copyright protected work. “It's been the case where you are allowed to use a portion of other people's work, recognizable chunks of other people's work, and use them in transformative cases.”

While there are prominent lawsuits, like the New York Times vs. OpenAI and 8 other dailies suing OpenAI and Microsoft, she sees others being dismissed. Laura believes it’s because those cases met three thresholds, “It was for research purposes in the first place and has resulted in the generation of new technology. It created works that are nothing like the original, and it uses a combination of comprehension and collage.”

For Aya, her artistic and business ethos is that transparency and ethical data practices are crucial. For others in the same camp, she highly recommends hiring an ethical AI advisor to partner in ensuring ethical compliance. “We are actively thinking through the quality of our data. Who is involved with it? How do we provide that recognition or acknowledgement to the data that we're using?”

This led the ImmerseMe.ai team to actively train their AI model on data they own in order to create their visuals, approaching it with a 50 question process called Datasheets for Datasets (a key framework LA Tech4Good teaches in our data equity workshops). “I have an artist’s perspective. You're still collecting all your information. You're not fully using what's out there. There’s still some level of intention and responsibility that goes into the digital artwork.”

How do we marry the cool & the scary? Using AI in inspiring & ethical ways

Her work in publishing has led Laura to cultivate friendships with best-selling authors, many of whom have come to her with the worry that their work is being used to make someone else money. She sees it another way.

“We all want to support science. If you, a master of words, were asked to use your work to democratize programming in order to build a more natural language understanding of people, I think the vast majority of people would say ‘Yes.’ It's very startling to be informed without consent that your works have been fed into a machine in their entirety. But, the ethics of AI cannot be weighed just by looking at any individual industry,” pointing out that in one area of medicine, two antibodies have been discovered through the use of AI in just the last ten months, something that hasn’t happened in fifty years.

Aya believes that progress in AI can benefit through policies of transparency, communication and education, referring back to those series of questions. “As we're designing our AI systems, what are the unintended consequences? What are our blind spots? What are our principles? What are our values when it comes to AI? And then as we're designing all of our products, we're actively thinking, how long should we store people's data?”

Aya also emphasizes the importance of communication feedback loops to non-technical audiences and users of their product. “I’m going against this black box mentality,” where companies hide how data is obtained and used to train models. By leading with transparency, “We get a lot of valuable insights this way. It’s been really helpful in shaping our direction.”

Moving forward, Laura's company has publishing-related patents and projects in the works, including an algorithm to address the nuances of language that have not been met by current LLMs. Another is an AI tool that holds promise for addressing the bias of human gatekeepers in the publishing and entertainment industries. Focused on handling the “slush pile” their AI prioritizes promising book manuscripts or film/TV scripts, providing publishers or studios with data-driven insights to make informed decisions, making it easier to find and evaluate high-quality scripts, leveling the playing field for all creators.

Aya is continuing her founder’s journey with Creative Aya and ImmerseMe, the intersection of AI, natural language processing, afro futurism and art, along with her senior role as a data scientist at Deloitte.


The formal part of the evening was closed out with a moment for Palestine, with the poem If I must die, let it be a tale by Refaat Alareer, beloved writer, poet, professor, and activist from the Gaza Strip, read by Karen Borchgrevink of LA Tech4Good.


About the author & speakers

Kimberly Owens

Kimberly Owens (she/her/ella) is a multidisciplinary professional, writer, entrepreneur, and founder of Buzz Boulevard, delivering bespoke content strategy and event marketing services to B2B and nonprofit clients in a variety of market sectors. Her decades in journalism and devotion to social justice is intertwined with her advocacy for responsible AI, volunteering her time to the anti-recidivism program, Re-Imagine, along with LATech4Good, and All Tech is Human.

Faith Umoh

Faith 'Aya' Umoh, the founder of Creative Aya and ImmerseMe AI, stands at the nexus of co-creativity, culture, and AI. Her journey into the digital art world was sparked by her research at Stanford University, where she worked on a Non-Parallel Multimodality Co-creative AI system, setting the stage for her innovative approach. With over 14 years of data science experience, including a significant tenure at Fortunate500 companies, Aya is recognized by Andreessen Horowitz as the Future of Digital Art. She explores the intricate connections between nature, interactive technology, and human connectivity. Her studio's work has been experienced in 7 countries and 12 showcases, including Intereality at Bitform Gallery and the Max Ernst Museum's inaugural digital exhibit, Surreal Futures. Aya is committed to using AI as a tool for fostering connection and exploring new dimensions of creativity, nature, and culture.

Laura Randant Crenshaw

Laura Randant Crenshaw is the Founder and CEO of Mythulu. Through this innovative company, Laura pioneers tools using cards to invigorate the artist's writing process, recognizing the human brain as an imitation machine and offering stimuli to catalyze creative breakthroughs beyond comfort zones. She uses storytelling to bridge diverse fields including business, entertainment, medicine, and wilderness exploration. With a fervent interest in the economics of art and a knack for narrative, Laura is also an accomplished author.

From Women in Big Data SoCal

Odessa Reyno, Director of Women in Big Data SoCal chapter stated that night, “We believe in the power of connection and collaboration to drive positive change. What sets Women in Big Data apart is our commitment to bridging the gap between global resources and local need. Today’s event is a testament to that mission.”

“As we delve into the world of generative AI for creative workflows, we're not just exploring the latest technological developments. We're also shining a spotlight on the unique perspectives, talents, and insights that women bring to the table. And by showcasing the expertise of our esteemed speakers, we're not only advancing the conversation on generative AI, but also highlighting the incredible contributions of women in this rapidly evolving field,” Reyno concluded.

Previous
Previous

3 Digital spools of educational justice

Next
Next

AI is here for healthcare