Story perspectives

Google unveils free Gemini Omni Flash video AI

5/21/2026

26 5

Summary

Google launches Gemini

Google launched Gemini Omni and its video version Gemini Omni Flash at Google I/O on May 20, 2026.
Gemini Omni Flash lets users create videos from prompts and is free in the Gemini app, YouTube Shorts and YouTube Create.
Google adds a SynthID watermark to all generated media and cites DeepMind CEO Demis Hassabis, who said the model can turn prompts into scenes like a circle becoming a black hole.

Summary

Google integrates Gemini

Google has repositioned Gemini as an AI operating system tightly integrated with Gmail, Calendar, and Veo-powered video tools.
Gemini is free, fast, and already installed on every Android phone.
The Daily Brief and Gemini’s agents automatically scan emails and calendar events to draft replies, resolve conflicts, and book meetings.
Because Gemini lacks token or blockchain integration, privacy-centric decentralized AI projects retain a niche untouched by Google’s data-driven model.

Summary

Google Unveils Avatar

Google unveiled an “avatar” feature that inserts a digital clone of users into any AI-generated video.
The avatars run on the Omni Flash model and are accessible in Flow, the Gemini app, and YouTube.
Google positions the tool as a competitor to OpenAI’s now-defunct Sora app, aimed at social-first selfie deepfake creation.
Creators can save custom instructions and set automated workflows that automatically sort similarly styled clips into folders.

Summary

Google Deploys Detection

Google is rolling out AI-detection features to Chrome and Search to identify deepfake images.
SynthID verification is now integrated into Search tools Lens, AI Mode, and Circle.
Chrome will let users select an image and ask “Is this AI-generated?” via a side panel.
C2PA data is now verified alongside SynthID.
Google Cloud introduced an AI-detection API for the Gemini Enterprise Agent Platform, initially for trusted partners.

History

Historical Context

Google’s 2015 release of DeepDream opened public debate over synthetic images, prompting open-source tools and early detection concerns. The AI boom of the early 2020s, marked by OpenAI’s 2021 DALL-E and 2024 text-to-video model Sora, accelerated demand for authenticity safeguards. Today Gemini Omni Flash embeds SynthID watermarks and Chrome detection, extending DeepDream’s legacy.

1 of 5

Google launches Gemini

Google launched Gemini Omni and its video version Gemini Omni Flash at Google I/O on May 20, 2026.
Gemini Omni Flash lets users create videos from prompts and is free in the Gemini app, YouTube Shorts and YouTube Create.
Google adds a SynthID watermark to all generated media and cites DeepMind CEO Demis Hassabis, who said the model can turn prompts into scenes like a circle becoming a black hole.

1 / 5