At the Google I/O 2025 developer conference on Tuesday, Google introduced its cutting-edge video-generating AI model, Veo 3, which enhances the experience of video creation by adding audio elements. This innovation allows users to generate sound effects, background noise, and even dialogue in conjunction with the videos it produces, marking a significant advancement from its predecessor, Veo 2, which already set a high standard for video quality.
Veo 3 is accessible starting that day via Google’s Gemini chatbot app exclusively for subscribers of the $249.99-per-month AI Ultra plan. Users can initiate video creation by providing prompts in text or image formats. Demis Hassabis, the CEO of Google DeepMind, highlighted that users can describe characters and settings, as well as suggest dialogue and its tone.
The landscape of video-generating tools has become crowded, with numerous startups—including Runway, Lightricks, and Pika—competing alongside tech giants like OpenAI and Alibaba. In this saturated market, Veo 3’s ability to sync generated sounds with its video content could be a key differentiator, assuming Google realises its ambitious promises. Although AI-generated sound tools are not new, Veo 3’s unique capability to interpret raw video pixels for audio synchronisation could set it apart from the competition.
Notably, the development of Veo 3 builds upon DeepMind’s previous innovations in “video-to-audio” technology, which aimed to generate soundtracks by training models on a diverse range of sounds and corresponding video clips. While specifics on the training data remain undisclosed, it is widely speculated that YouTube—owned by Google—was likely a significant resource.
To address concerns related to the spread of deepfake content, DeepMind has integrated its proprietary watermarking technology, SynthID, to embed invisible markers in the frames produced by Veo 3.
Despite the potential benefits of such tools, many artists express apprehension, given the disruptive impact they may have on the creative industries. A 2024 study commissioned by the Animation Guild estimated that over 100,000 jobs in the U.S. film, television, and animation sectors could be affected by AI advancements by 2026.
Additionally, Google unveiled enhancements for Veo 2, introducing features that allow users to input images of characters and scenes for improved coherence, understand complex camera movements, and modify video objects or framing—all of which will soon be available on the Vertex AI API platform.
In summary, Google’s Veo 3 represents a significant leap in AI-driven video content creation, bringing audio capabilities into the mix and setting the stage for new possibilities in dynamic media production. However, the implications of such advancements raise valid concerns within the artistic community regarding their future.
Fanpage: TechArena.au
Watch more about AI – Artificial Intelligence
