Google

Unleashing the Potential of Google’s Revolutionary AI Technology

Published

1 hour ago

May 23, 2026

Google’s new anything-to-anything AI model is wild

Last Year’s Experiment with Deepfaking Stuffed Animals

Last year, I decided to have a little fun by deepfaking my child’s stuffed animal to create the illusion that his plush deer was enjoying a vacation.

Exploring the World of Generative AI with Google’s Omni

It was an interesting experiment inspired by a Gemini ad run by Google. The experience led me to ponder the fine line between harmless generative AI fun and something more concerning. The capabilities of the tools for creating realistic videos are remarkably advanced, requiring minimal effort and expertise. This trend is further evolving with Gemini’s Omni era.

Introducing Omni: A New Frontier in Generative Models

Omni is a new family of generative models that aims to transform various inputs such as photos, videos, and text into diverse outputs. Initially focused on video creation, Omni Flash is the first model released by Google in their AI video generation and editing platform, Flow. While the previous model, Veo, is still available, Omni offers several enhancements.

Testing Omni’s Capabilities with AI Buddy

With Omni, users can upload a video and pair it with a text prompt to initiate AI-generated content creation. Google claims that Omni integrates more real-world knowledge into video production, enhancing character consistency throughout the content. To verify these claims, I resurrected AI Buddy for another adventure.

The Mixed Results of Omni’s Creations

The outcomes of my experiments with Omni were a blend of impressive and perplexing. While some videos demonstrated improved consistency compared to Veo, there were still occasional AI glitches, such as sudden orientation changes during skydiving scenes.

Artistic Freedom and Quirks in AI-Generated Videos

Granting Omni artistic freedom resulted in a whimsical video where Buddy packed a jar of honey for a tropical vacation. However, the inconsistencies in the appearance of the honey jar throughout the clip showcased the limitations of the model.

Challenges and Costs of AI Video Generation

While text-based prompts can guide video edits, the process incurs costs ranging from 15 to 40 credits based on the scene’s length and complexity. Each round of edits consumes 40 credits, highlighting the financial investment required for generating content with Omni.

I can genuinely say I wasn’t prepared for what I saw

Deepfaking Personal Videos with Omni

Exploring Omni’s ability to add AI-generated elements to real videos, I transitioned from AI Buddy to deepfaking myself. The results were surprisingly convincing, with minor AI tells like manufactured sounds and duplicated background characters.

We’re definitely deep in the uncanny valley

Reflections on the Evolving Landscape of AI Technology

The rapid advancements in AI technology, as exemplified by tools like Omni, continue to blur the lines between reality and simulation. While the process of creating AI-generated content has become more accessible, the nuances and imperfections remind us of the ongoing journey into the uncanny valley.

All images and videos in this story were generated by Google Gemini.