# OpenAI Releases DALL·E 3 System Card Detailing Safety Improvements

OpenAI has published the official system card for DALL·E 3, its latest AI image generation model, providing transparency into the system's capabilities, limitations, and safety measures.

The system card reveals significant improvements in content policy enforcement compared to previous versions. DALL·E 3 includes enhanced filters to prevent the generation of violent, adult, or hateful content, as well as stronger protections against creating images of public figures or copying the styles of living artists without permission.

Key technical details show that the model was trained with improved safety datasets and human feedback to reduce harmful outputs. OpenAI implemented new safeguards to decline requests that violate usage policies, including attempts to generate misleading or deceptive content.

The document also outlines the model's integration with ChatGPT, which acts as an additional safety layer by refining user