While video models often “hallucinated” unwanted details – for example, additional fingers or unexpected objects – VEO 2 produces these less often, which makes more realistic outputs.
Our commitment to safety and responsible development was led by VEO 2. We were deliberately measured in the growing availability of VEO, so that we could help identify, understand and improve the quality of the model, while slowly introducing it via videoofx, YouTube and Vertex AI.
Like the rest of our image and video generation models, VEO 2 outputs include an invisible synthesis watermark that helps to identify them as generated by AI, helping to reduce the chances of disinformation and improper determination.
Today we are introducing our new VEO 2 capabilities to our Google Labs video generation tool, Videofx and increase the number of users who can access it. Visit Google Labs To sign up for the waiting list. We also plan to expand VEO 2 to shorts on YouTube and other products next year.
