Google's Veo 3 AI Generates Videos with Sound

Google has unveiled Veo 3, the latest version of its AI video generation model. This new AI can create accompanying audio, including sound effects, background noise, and even dialogue.

Announced at Google I/O 2025, Veo 3 is available now in the Gemini chatbot app for Google AI Ultra subscribers ($249.99/month). Users can prompt Veo 3 with text or images to generate videos with sound.

Veo 3: Entering the Era of AI Video with Sound

Demis Hassabis, CEO of Google DeepMind, highlighted Veo 3's ability to create a complete audio-visual experience. Users can provide prompts describing characters, environments, and desired dialogue, including specific sound descriptions.

While the AI video generation field is becoming increasingly crowded with startups like Runway, Lightricks, Genmo, Pika, Higgsfield, Kling, and Luma, as well as tech giants like OpenAI and Alibaba, Veo 3's integrated audio generation offers a significant advantage.

Although AI sound generation tools exist, Veo 3's unique ability to analyze video pixels and automatically synchronize generated sounds sets it apart. This feature streamlines the video creation process and enhances realism.

cooking up something tasty for tomorrow… pic.twitter.com/wyIRMsXkFG

— Demis Hassabis (@demishassabis) May 19, 2025

Veo 3 likely builds upon DeepMind's previous work in "video-to-audio" AI. Google owns YouTube, and it's possible that YouTube content was used in Veo 3's training data, although DeepMind has not confirmed the specific sources.

Addressing Deepfake Concerns

To mitigate deepfake risks, Veo 3 uses Google's SynthID watermarking technology to embed invisible markers into generated video frames.

Impact on the Creative Industry

While Google promotes Veo 3 as a creative tool, its potential impact on the job market is a concern. A 2024 study suggests that AI could disrupt over 100,000 film, television, and animation jobs in the US by 2026.

New Veo 2 Features

Google also announced updates to Veo 2, including improved control over character design, scene composition, and camera movements. These features will be available on the Vertex AI API platform soon.