Google's new Omni model can transform any input into anything else, from stuffed animals to Eiffel Tower selfies. The results are both impressive and unsettling, with some clips hitting the uncanny valley harder than others.

The model requires credits for use, which can quickly add up, especially when making edits. While it excels at consistency in video characters, inconsistencies emerge where AI jumps into the narrative.

For deepfakes, the tool is surprisingly convincing. A husband tested one clip and was fooled by a fake pasta-eating scene, underscoring just how realistic these videos can be.

The future of AI-generated content looks bright but costly. As tools like this become more accessible, the line between reality and fiction will blur further, prompting new discussions on authenticity in media.