Google revealed Gemini Omni on Tuesday at its annual I/O developer conference, positioning the new service as the next leap in synthetic media. The multimodal AI can ingest virtually any combination of text, images or video clips and output realistic, high‑fidelity video. Powered by the Gemini modeling architecture, Omni treats each input as part of a unified world model, allowing the system to maintain consistent visual and physical logic across the generated footage.
At launch, users will create videos using text prompts, still images or existing video footage. Image‑to‑video and text‑to‑video functions are available now; a future update will add pure text generation capabilities. The tool also doubles as an editor. After generating a clip, creators can feed the result back into Omni, issue a new prompt and have the system replace or modify specific elements—changing a background, swapping a subject’s attire or even inserting a custom avatar that mimics the user’s voice and appearance.
Google has built safeguards into the workflow. Every Omni output carries a SynthID watermark that identifies the content as AI‑generated, a measure meant to curb the spread of deceptive media. The company says the watermark is automatically applied and cannot be removed by the user.
Access to Gemini Omni will roll out across several Google products. Paid subscribers can already experiment with the tool inside the redesigned Gemini app, where templates can be added to a camera roll with a single tap. The same functionality will appear in Google Flow and on YouTube Shorts later this week. Developers and enterprise customers will receive API access in the coming weeks, opening the door to custom integrations and broader commercial use.
Omni is offered in two tiers. The initial Flash version, which is currently available, delivers fast generation for everyday users. Google promises a more powerful Pro variant in the future, though it has not disclosed a timeline. By combining multimodal input, advanced physics simulation and built‑in editing, Gemini Omni aims to set a new standard for AI‑driven video creation while navigating the ethical challenges that accompany realistic synthetic media.
Dieser Artikel wurde mit Unterstützung von KI verfasst.
News Factory SEO hilft Ihnen, Nachrichteninhalte für Ihre Website zu automatisieren.