Trust & Safety
As progress in AI continues, xAI remains committed to the safety of our models and products. Advances in reasoning raise new risks — and unlock new layers of defense.
We integrate safety evaluations throughout the model development lifecycle, from pre-training through deployment.
We publish model cards, safety evaluations, and our Frontier AI Framework to keep the community informed.
We actively hire safety researchers and engage with the broader AI safety community to improve our practices.
Documentation
We publish detailed model cards for each of our frontier models, covering capabilities, limitations, safety evaluations, and intended use.
Reporting
If you encounter safety issues with any of our models or products, please reach out. We take every report seriously.
