<ul><li>Google has announced a preview of Veo 2, its new AI video model that understands the language of cinematography and has a better understanding of real-world physics and human movement.</li><li>Veo 2 aims to address the challenges of correctly modeling humans in motion and reduce the occurrence of artifacts in generated videos.</li><li>Google is also introducing improvements to Imagen 3, its text-to-image model, which generates brighter and better-composed images and follows prompts more faithfully.</li><li>Veo 2 will gradually roll out to Google Labs users in the US, while the enhancements to Imagen 3 are available to Google Labs users in over 100 countries.</li></ul>

Google's new AI video model sucks less at physics

Discover more