With a Data Licensing Framework in Play, Rights Holders Can Embrace AI 

In the past year, we’ve witnessed a surge in partnerships between content creators and AI developers. From text and images to videos and music, rights holders are increasingly licensing their data to startups and Big Tech companies to train AI models.

In the music industry, for example, some companies have successfully licensed music catalogs from rights holders to train AI music models capable of composing original music — demonstrating the potential for mutually beneficial partnerships between rights holders and AI developers.

However, there remain obstacles that must be resolved. Ongoing legal battles between major players in the industry highlight the importance of establishing clear licensing agreements and respecting intellectual property rights.

As AI continues to advance at a rapid-fire pace, it’s crucial for rights holders and AI developers to navigate these complexities, find common ground and work together to solidify the framework for AI data licensing that benefits all parties.

Key Considerations for Licensing Deals
It’s important to acknowledge that AI developers face significant costs in building and maintaining their models. For this reason, a purely royalty-based licensing model may not always be economically viable for every AI use case.

The process of creating an AI model involves two main stages that require considerable investments: training and inference. In the training stage, AI developers require substantial computational resources to feed vast amounts of data into their models, allowing the AI to continually learn and improve its performance. This process can be time-consuming and expensive, often requiring powerful hardware and large amounts of energy.

Once the AI model is trained, it enters the inference stage, where it is deployed to generate new content or perform specific tasks. While the computational costs of inference are generally lower than those of training, they can still be substantial, especially for applications being delivered to a large number of users for an extended period of time.

Given this outlay, finding a balance that recognizes the costs incurred by AI developers while fairly compensating rights holders is essential.

Benefits of Hybrid Licensing Models
The key to successful AI data licensing lies in adopting a hybrid model that considers the specific use case as well as the unique needs of both parties.

Hybrid licensing models that combine upfront fees with revenue sharing can strike this balance, ensuring both parties can benefit from the AI-generated content or services. In some instances, a flat fee licensing structure may be appropriate, particularly for onetime or limited usage.

This model provides rights holders with upfront compensation and allows AI developers to access high-quality data without ongoing financial obligations. However, for more extensive or commercial applications, a revenue share model can be a mutually beneficial approach.

Exclusive: What Media & Entertainment Thinks About Using Gen AI

Under this arrangement, rights holders receive a percentage of the revenue generated by the AI-powered products or services that utilize their licensed data, similar to a royalty pool utilized by streaming services.

The model aligns the interests of both parties, incentivizing rights holders to provide high-quality human-made data on an ongoing basis while paving the way for AI developers to sustainably monetize their offerings.

The Path Forward 
In the face of ongoing legal battles and court rulings surrounding AI and intellectual property, it’s important for rights holders to consider proactive licensing strategies, as waiting for the courts to establish precedents may leave them behind given the rapid pace of AI advancements.

By actively engaging in agreements and partnerships with AI developers, rights holders can shape the future of AI data licensing and see that their interests are protected.

VIP+ Explores Gen AI From All Angles — Pick a Story

Moreover, clear licensing terms and transparent attribution mechanisms are vital components of a successful AI data licensing framework. Creatives need visibility into how their licensed data is being used, and, when possible, AI-generated content should properly credit the original creators.

This approach to licensing must become a best practice in order to foster trust and collaboration between rights holders and AI developers. The key to adapting to the evolving landscape is in collaboration, in rights holders and AI developers working together to refine licensing models, establish industry standards and address the unique challenges posed by this transformative tech.  

Alex Bestall is the founder and CEO of Rightsify and Global Copyright Exchange (GSX), two companies standing at the forefront of the AI music revolution. 

Now dig into a VIP+ subscriber report ...