Google launches AI-powered video model Veo 3.1 with new tools

Google has launched Veo 3.1, the latest version of its AI-powered video generation model, adding enhanced audio features and more editing controls for users.

Veo is used within Flow, Google’s AI filmmaking tool, which has generated over 275 million videos since its release five months ago.

The update introduces richer, generated audio, and new options to edit video clips, such as inserting new elements and removing unwanted objects.

Users can now add or refine audio across features like “Ingredients to Video,” “Frames to Video,” and “Extend,” increasing creative control in scene and shot development.

Veo 3.1 also improves realism and prompt adherence, delivering higher audiovisual quality and true-to-life textures.

The model is available to developers via the Gemini API, to enterprise users through Vertex AI, and in the Gemini app. Some new editing tools, including object removal, will be rolled out soon.

🔗 Source: Google

🧠 Food for thought

Implications, context, and why it matters.

IP ownership and liability protection remain undefined for Veo 3.1 users

Veo appears on Google’s Generative AI Indemnified Services list 1 (a program that offers legal indemnity for certain products). Terms for version 3.1 commercial use and Intellectual Property (IP) ownership remain unclear.
Veo 3 runs under Pre-GA (General Availability) Offerings Terms as a preview 2. Companies can use it commercially with possible limits on support or compatibility.
The program covers third-party IP claims in general. The scope for Veo 3.1 video and audio needs a close read of the Google Cloud Service Specific Terms (the contractual documents governing product use).
Enterprises face added risk on ownership and liability until terms align with mature AI services.

Content provenance infrastructure will matter as AI video generation scales

Videos made with Veo on VideoFX carry SynthID watermarks 3. This is Google’s invisible, pixel-level watermarking technology embedded across every frame.
Detection needs voluntary upload to Google’s portal, and heavy edits can weaken the mark 4.
Startups and middleware can ship automated compliance layers. These blend SynthID checks with Coalition for Content Provenance and Authenticity (C2PA) standards 5 plus disclosure flows.
Enterprises and platforms that host user videos will need verification systems to manage liability and keep trust as AI clips look indistinguishable from authentic footage.