Our Mission & Approach to Safety
Powerful AI requires thoughtful development. As an open-source project, we are committed to building safe and beneficial text-to-video technology through community collaboration.
A Collaborative & Open Approach
We are engaging with researchers, artists, and domain experts to understand the potential risks and benefits of this technology. Our development process includes several critical safety steps:
- Building Robust Classifiers: We are developing tools to detect misleading content, hateful imagery, and other potential harms.
- Community-Driven Red Teaming: Leveraging the open-source community to test the model for vulnerabilities, biases, and harmful generation capabilities.
- Adhering to Best Practices: Following established safety protocols and contributing our findings back to the broader AI research community.