Today’s consumers are more discerning, scrutinizing product origins and demanding sustainability and ethical practices from brands. They demand more than just a product — they want to understand how products impact everything from the environment to our society. This same sensibility has already begun to enter the realm of generative AI content, and it’s poised to be a dominant narrative in 2024. While the technology promise is immense, it cannot crack one immutable truth: Where consumers go, brands follow. I expect to see a surge in the demand for the ethical sourcing of AI-generated content across branding, marketing and media.
Anxiety over deepfakes and the blurred lines of reality
AI-generated content has captured our imagination, able to generate photo-realistic images and cinematic videos with just a few keystrokes. And as this content proliferates, we both dream about the benefits it could unlock and worry about the harm it could inflict.
Already this year we have heard about celebrity deepfakes and AI-generated robocalls in the presidential primaries, highlighting the new challenges that accompany generative content as it rises to prominence. While these may be newer concerns for the general public, creators, artists, photographers and videographers have expressed unease about the potential for abuse even before the technology made it to the mainstream.
The existential threat to creatives and watering down of creativity
Humans have been the driving force behind artistic expression for tens of thousands of years. Beginning with oral storytelling and music, evolving to literature and paintings, and more recently with film and photography, we’ve created a window into nearly all of human history. As our lives have become increasingly digital, these artistic breadcrumbs have also created a massive data footprint — all due to human ingenuity. Generative models have benefited from this incredible data footprint, capable of producing stunning visuals of the world — both real and imagined. Yet none of these visuals are originals. Every piece of generative content is a derivative of something else — specifically, an object humans created.
Without protections to support creators, AI content will merely mix derivatives. Not only do these derivative works potentially offer a bleak view of human ingenuity and creativity, but experts say models trained solely on AI-generated content suffer from “model collapse,” forgetting what they initially learned from human-created data and generating more errors over time. Avoiding this fate requires ongoing access to fresh, human-originated data. Absent strong protections for IP rights and creators’ ability to earn, we’ll lose access to both the art that has been central to the human condition, as well as the data that has fueled the initial generative boom.
What is ethically sourced content?
Just as more consumers want to know the coffee they buy is ethically grown, harvested and packaged — and the farmers and other workers are fairly compensated — we’re more interested than ever in the origins and chain of custody of what we buy. It doesn’t feel good to think you’re contributing to unpleasant working or living conditions for another person. In the same way, no one wants to think of themselves as negatively impacting an artist’s livelihood.
Artists are asking these same questions. A survey of Storyblocks’ current contributors found that many acknowledge AI will play a role in the future of visual media, and they view this change positively. But they want to understand more about how they’ll be compensated for these new initiatives, how their rights as artists could be affected and where copyright fits in. Given the uncertainty surrounding several major intellectual property lawsuits, many of these artists also voiced concerns about AI’s long-term impact on their businesses.
As generative AI collides with IP rights and a more socially aware consumer base, I believe there will be growing demands for “ethically sourced” AI labeling that discloses the details of the work’s genesis. Brands that embrace this trend will have an opportunity to win trust and loyalty, while also helping to create a more sustainable creative economy as we transition to the generative age.
I believe that “ethically sourced” AI content will:
- Protect original copyright holders by compensating them directly for their work.
- Offer transparency to customers about the origins — both the art and the rights — of the work used to train generative models.
- Provide clear labeling so that when content is distributed across the web, consumers and brands can easily differentiate between AI- and human-generated material.
While I expect consumers of all shapes and sizes to demand this level of transparency, B2B customers and brands will have an even greater incentive to understand the origins of generative content and seek guarantees that any derivative works will be squarely protected by appropriate usage and distribution rights. Without these guarantees, companies will be at risk. Already we see lawsuits aimed at companies training large language models. When businesses whose core expertise is not generative AI leverage the technology to improve their own offerings, they want to ensure they are protected.
What comes next for generative AI content
We’re already seeing movement from major players in the space that back up these predictions. In February, Meta announced it would begin labeling AI-generated photos on its platforms and penalizing users who fail to disclose the use of AI in creating realistic videos or audio. Meta also plans to watermark images from its own and others’ AI tools and is developing ways to detect disguised AI content. Meta — and other companies — should establish a framework to prevent the use of their models (and the artists’ data) to create misleading or illicit images. And just two days after Meta’s announcement, Google and Microsoft followed suit, joining an effort by the Coalition for Content Provenance and Authenticity to label AI-generated content. The coalition likens these credentials to a “nutrition label” connecting the ingredients and recipe to a piece of content.
I see a future where organizations facilitating the creation of AI-generated content will need to devote time and attention to developing that ethically sourced label, offering what amounts to a public guarantee that original artists are being fairly compensated and customers can create generative works without creating future legal or business risk. Just as businesses dealing in content creation include these features as part of their product, companies using generative AI to create media will need to demonstrate a similar level of provenance for the data used to power their models.
Technology cycles tend to follow a similar pattern. In the first phase, technology takes the lead as individuals and businesses explore the new frontier of what is possible. This is what brought us ChatGPT and other similar LLMs. In the second phase, which we are entering now with generative AI, customers reassert control. Products focused on solving customer problems — genuinely improving workflows and outcomes — will reap rewards with adoption and the ability to monetize. Similarly, brands that show they are meeting customers where they are — in this case, ensuring that they are building a sustainable industry that continues to reward human creativity — will be rewarded with loyalty.
Providing transparency through “ethically sourced” labeling will enable companies and individuals to confirm they have full rights to use AI-generated content as intended without violating laws or ethics. As generative AI becomes increasingly mainstream, consumer interest in and demand for this kind of transparency will grow substantially, just as it has for so many other industries.
