The video AI model Veo 2 from Google received new powerful tools that enable users to create cinematic content and improve real video footage more easily. The camera direction presets of Veo 2 are currently accessible through Google Cloud’s Vertex AI preview mode. These presets enable users to direct shot composition, camera angles, and pacing through basic text instructions. The new video editing options, including drone-style POVs alongside panning shots and time-lapse effects, enable Hollywood-standard control for creators who lack video editing experience.
The latest Veo 2 editing tools feature inpainting for background removal and outpainting for AI-generated frame extension, which matches the original footage in a similar way to Adobe’s Generative Expand feature. Interpolation within the update enables users to generate intermediate frames that produce smooth transitions between static images.
Google enhanced its text-to-image model, Imagen 3, through automatic object removal capabilities, which produce more natural results. The tools allow L’Oreal and Kraft Heinz to produce content in hours instead of weeks, which previously took weeks.
The audio division of Google introduced Lyria as a text-to-music generator through private preview while adding “Instant Custom Voice” to Chirp 3, which generates authentic custom voices from 10-second audio samples. The new transcription tool detects individual speakers when processing conversations involving multiple people.
Google integrates these AI enhancements into Gemini 2.5 Flash and the speed-optimized model while updating their enterprise agentic AI tools to enable agents to work across platforms, including Salesforce and PayPal. Through its recent technological advancements, Google speeds up the development of generative content for videos, audio, and images.
