The Future of Commercial Creative Depends on What We Legislate and Appreciate

The last few years have been a major shakeup for creators and content users, all kicked off by two little words you’re probably all sick of hearing: generative AI. Tools like Dall-E 3, ChatGPT, and Suno can transform simple descriptions into images, text, and songs in mere moments. But the gold rush of trying to use these tools to completely replace the work of humans might be coming to a close as legal and ethical issues start to catch up to the fantasy of AI’s potential.

The legal future of generative AI

The technology behind generative AI currently outpaces legal regulations. Generative AI works by training models on gigantic libraries of labelled existing content. Because it’s legally ambiguous whether training on a copyrighted work constitutes copyright violation, gen AI companies are treating it as a free-for-all, training their models on everything and anything. This allows models to emulate the style of artists they trained on – the legality of which is also undecided.

Compounding this legal challenge is the difficulty of ever proving what any model is or isn’t trained on. The training libraries are so huge that combing through them to find specific works isn’t viable at any scale. But that could all change soon: a bill proposed by California congressman Adam Schiff would force gen AI companies to disclose any copyrighted material used to train any AI model that’s used commercially.

This bill passing would be a massive shakeup. It would radically change how AI models are trained, requiring better classification of training data, and as a result, likely smaller and less versatile training databases. If training on copyrighted data ultimately is ruled to be copyright violation, massive retraining would have to be done to avoid lawsuits.

On the one hand, the bill has support from powerful lobbying groups such as the RIAA, SAG-AFTRA, and other entertainment industry organisations. On the other, nearly all publicly traded big tech companies, including Alphabet, Microsoft, and Meta, have invested untold billions into AI and will eagerly fight back. Regardless of how this bill plays out, the battle over copyright and AI isn’t ending anytime soon. We encourage you to be cautious in this time of uncertainty, and not overcommit to your own AI usage. Beyond legal ambiguity, there are other issues and better alternatives.

What it means to rely on generative AI

The appeal of generative AI was being able to make something that fits your specific needs without spending time or money. Advocates hoped it would be as easy as grabbing an image from Shutterstock, but as specific as going out and taking a picture yourself. And sure, it can sort of accomplish this – you can get something that looks or sounds like what you want, no matter what that thing is. It’s useful for prototyping concepts, to help you lock in on what specifically you want. But when it comes time to publish, you shouldn’t rely on gen AI to get what you want.

For starters, using generative AI can open yourself up to unwanted criticism. Beyond quality issues, there’s plenty of other reasons audiences may respond poorly to you using generative AI content: viewing model training and style emulation as a form of plagiarism, concern around the environmental impact of the intensive computing necessary for generative AI, stress around AI replacing human jobs, and more.

Of course, there are forms of AI where none of these controversies apply. For example, Catch+Release's semantic search is enhanced by AI tech to make it more useful. AI has provided many tools to improve your products and productivity, and you owe it to yourself to investigate them. At the same time, you owe it to your users to apply AI in ways consistent with their values.

AI is a broad term and can include any tool designed to solve hard problems with specific techniques. You may find many helpful uses of AI that don’t have the issues of generative AI. AI can help you organise your media based on its content, rather than labels, like with our semantic search. It can help you prototype and iterate on creative ideas, letting you see a rough draft of different visions easily or add AI-enhanced edits. It fills in many workflow gaps where manual work would be too tedious or take too long, making the work of humans smoother and more effective. We encourage using AI where it works best, and keeping humans where they work best. We’ve found that for content creation, AI isn’t the best.

When we talk to customers, we find that AI just isn’t hitting the necessary quality threshold they expect from human-created content. AI material often has distracting flaws, or is uncannily generic. If your readers are wondering “is this real?”, they won’t be engaged with the content itself. And once they’re focused on that, the message of your content can feel contradictory. For example, if you’re writing about “high protein delicious meals”, and you use an AI generated picture of a meal, your readers will be sceptical about any recipe you post. After all, why wouldn’t you just use a picture of the actual dishes if they were so good?

What this boils down to is that AI-produced content is inherently inauthentic. Because it reflects an amalgamation of the whole internet’s conception of a concept at the moment of training, it will always lag behind the most dynamic and most individual perspectives. Your audiences will never see it as something fresh or exciting, it will always be reiterating what’s already out there. Your campaigns and your readers deserve the cutting edge.

The power of user-created curated content

The best way to get authenticity, specificity, and value alignment without sacrificing speed and cost-effectiveness is user-created content. These days, billions of people are uploading pictures and videos giving glimpses of every aspect of their lives. This content is as authentic as it gets: real people with real experiences, on the forward edge of every cultural moment. In fact, it’s user-created content that creates the cultural moment itself. You can’t get more relevant. And with such a depth and variety of content out there, there’s always something that fits your exact needs.

One potential roadblock for user-created content in a commercial context is licensing. The users creating this content need to consent to the content’s use, be credited for their role, and compensated for the usage. As this content could come from any individual anywhere, it can be challenging to set up these relationships. That’s where Catch+Release comes in.

At Catch+Release, we’ve built a community of content from creators eager to license their work for your campaigns. You can dive into this huge library with cutting-edge search technology and find exactly what you want. And if you find the ideal content somewhere else, Catch+Release can help you arrange for proper licensing so you can use it in your campaigns.

It’s our mission to make the commercial media of the world more vibrant by bringing individual experience into the spotlight. We want to make that easy and mutually beneficial for marketers and creators by making sure creators consent to every usage of their content, and that every usage is credited and compensated. And we want marketers to be empowered to find the content that best fits their needs and help them enjoy the benefits of truly authentic and engaging content.

We don’t know exactly what the future holds for generative AI, but we’re sure user-created content is here to stay. Start exploring our Creator Community Marketplace to see how to make your next campaign authentic while staying on time and under budget!