Verus

For nearly a decade, BCGuardian has supported rights owners to ensure fair compensation for creators’ efforts. The challenges we face have evolved, but the objectives have largely remained the same – limit the visibility and reach of unauthorized content on the web.

Achieving these objectives relies, to varying extents, on a process of notice-and-takedown. However, the most urgent content protection challenge today is not one to be fought under the existing notice-and-takedown paradigm.

The rapid pace of AI development, particularly Large Language Models (LLMs), is surpassing rights owners’ current content protection strategies. LLMs are training on rights owners’ protected content, raising critical challenges which cannot be addressed through an outdated model of detection and enforcement.

LLMs are training on rights owners’ protected content, raising critical challenges which cannot be addressed through an outdated model of detection and enforcement.

This is why we created Verus.

Verus is a first-of-its-kind AI Content Intelligence Platform. It provides rights owners insight into the use of their content by nearly a dozen of the market’s leading LLMs. At the core of Verus is a proprietary interrogation algorithm which samples content from authorized works to determine the likelihood of inclusion within a model’s training data.

Verus also offers features for Retrieval Augmented Generation (RAG) enabled models and for model output matching. Our module for RAG-enabled models facilitates the detection of content retrieved from online sources which is synthesized into responses. The output module detects reproduction of content in LLM responses. Such output can be generated following prompting of models by Verus or delivered by rights owners through our BYOO (“Bring Your Own Output”) feature.

Verus offers rights owners the opportunity to take a significant step toward solving the challenge of uncompensated use of copyrighted content in AI training.

Verus achieves this by providing rights owners with quantifiable and replicable data related to the use of their content within LLMs. It provides the evidentiary support required for audits of licensed content or for legal escalation related to unlicensed content. Verus enables advocacy by rights owners and trade organizations by providing data to support industry standards. It enables rights owners to signal transparently to internal and external stakeholders about their content’s exposure to unauthorized training. And most importantly, Verus encourages ethical practices in LLM training by empowering rights owners to ensure their intellectual property rights are respected.

Verus encourages ethical practices in LLM training by empowering rights owners to ensure their intellectual property rights are respected.

We believe in adequate compensation for rights owners’ contributions to the tremendous potential AI unlocks. Verus is the tool to provide rights owners the intelligence required to navigate the intersection of AI innovation and copyright protection.

We’re thrilled to debut Verus, and hope you will be a part of the story.