Navigating Rights Management in Generative AI

Nov 6, 2023

Intro

With more and more brands becoming aware that Generative AI technologies exist, many are looking to better understand how the rights work on using these assets for marketing materials. Often times, the barrier to adopting Generative AI comes from legal barriers and not understanding the upfront implications of using imagery that has been generated by AI. Marketers have questions like:

  • Who owns the content at the end of the day?

  • What sort of license terms and rights exist for AI generated images?

  • Do model releases and property releases apply in AI generated imagery?

In this article, we will discuss the major considerations of rights management in Generative AI for enterprise marketers such as ones working at CPGs, pharmaceuticals, or financial services sectors.

Section 1 - Intellectual Property / Ownership

Ownership of AI generated images is a complex and evolving legal area. However, the general rule of thumb is that copyright law attributes authorship to "human creators." Therefore, since AI tools are prompted by humans, the humans own the copyright to the works produced by AI. This might be called "Human Creator Ownership" in some jurisdictions. There are instances where an employee or a contracted entity is creating works using AI on behalf of an organization - in this case, we would consider this a "Work-for-Hire Agreement" where the employer or the commissioning party would be the owner (unless there was a specific agreement signed that defined the working engagement).

For example, with Midjourney, a popular AI image generation tool (as of Nov 2023), their subscribers own all the images they’ve created, even if their subscription has expired, and they’re free to use those images however they’d like. If the business grosses more than $1M USD a year, you just need a Pro plan or above to use those images commercially. This makes it extremely viable for businesses to start using AI generated images from an IP perspective since ownership is effectively baked into the process.

Section 2 - Usage Rights (i.e. Term, Global vs Local, Channels)

Traditionally, photography and imagery were produced either through photoshoots (where you have to negotiate the usage rights) or through stock photography marketplaces (where you can choose to buy out the image fully or license it for a fixed term). Both of these options usually have fixed term as a default, whereby we have seen many brands struggle to handle rights for all of their image assets.

There are 3 major considerations when it comes to usage rights:

  • Image Term: How long can I use the images?

  • Global vs Local: Are there regional jurisdictions that are restricted?

  • Channels: Am I limited to 1 channel or can I use them across any channel?

The beauty of AI-generated imagery is that those 3 considerations that often need to be negotiated get completely wiped out because images produced by AI are default to be perpetual across any channel globally.

At AeroBoard, we offer our clients an unlimited, global, exclusive license to use across any channel in perpetuity - translating to effectively no restrictions.

Section 3 - Cost of Model and Property Releases

One of the biggest cost drivers with traditional photography and full production shoots are the exorbitant costs associated with not only acquiring models and properties, but obtaining releases. Let's say you were trying to photograph a model posing as a doctor in specific hospital setting… you would need to consider some of the following costs:

  • Negotiating and paying for the Model's time and likeness

  • Negotiating and paying for the rights to use the Model's face and body for X term

  • Negotiating and paying for the Property's time and likeness

  • Negotiating and paying for the rights to use the Property for X term

All of those costs go away with AI generated imagery because we are creating fictional scenarios based on examples of real life things that the AI has been trained on. No longer do we need to negotiate a model's fee because the AI generated model is not a real person (but does look like one!).

Real-Life Example: One of our Pharma clients had an agency shoot a bunch of content using a famous comedian as a model but those rights were on a fixed term. When those rights expired, they were quoted several $100k USD to continue to use those same images. The alternative was to create brand new images using AeroBoard which totally mitigates the need for negotiating/paying for model and property releases entirely. This was a way better option for the client than renewing rights or going to re-shoot through an agency.

Conclusion

As enterprise marketers start navigating the waters of Generative AI, rights management is a consistent conversation starter. There are several considerations from an IP, usage rights, as well as releases standpoint. AI-generated images have so much benefit to traditional means of sourcing imagery because of how much complexity it reduces during the production process. AeroBoard recognizes our responsibility of ensuring that the creative we produce for our clients upholds to the standards of your organization and provides flexibility for you to use these assets for a long term. Get in touch with us to learn more about how we can help you create bespoke, branded content using AI.