Cullen International’s Global Trends service dissects one of the most pressing issues in the tech space today. It does so in a new report titled When models collide: how copyright could shut down AI’s supply of free data. The report takes a closer look at the legal disputes that have emerged over training AI models with copyright-protected content. This very content is the lifeblood of many business models at the heart of the media sector. The report discusses several key aspects of this debate, such as:

• The current state of the global clash between the media and AI sectors over copyright and data.
• Whether developing AI with data mined from the internet infringes copyright and, if so, if any exception applies.
• If no exception applies, how should creators be compensated for past AI training? And how should they be compensated in the future? 

The report has global coverage

It contains a breakdown of international copyright protections under the Berne Convention international agreement and a brief explanation of why AI needs third-party data to begin with.

It also explores the legal arguments against and for upholding copyright protections for data extracted from the internet. Additionally, it includes a survey of the applicable law in several jurisdictions in the Americas, Asia-Pacific, and Europe.

Potential resolutions

The report discusses several potential resolutions to these disputes. There are several paths forward. For instance, some AI developers are voluntarily striking licensing agreements with different media organisations, including social media platforms, for their data-rich content.

In one particular case, the report focuses on how much Open AI has reportedly paid some publishers to license their content for AI model training.

For more information and access to the full report, please click on “Access the full content” - or on “Request Access”, in case you are not subscribed to our Global Trends service.