Google AI transforms photos into videos on Honor phones, though amusing errors persist

Honor debuts Google’s Veo 2 image-to-video AI on smartphones before the technology reaches Gemini users, offering a glimpse into how AI video generation is becoming more accessible. This notable early implementation showcases both the potential and current limitations of consumer-facing AI video tools, while establishing a precedent for how similar features might be monetized as they become more widely available.

The big picture: Chinese phone manufacturer Honor has integrated Google’s Veo 2 image-to-video AI generator into its latest smartphones before the technology is widely available to Google’s own Gemini users.

• The feature will debut on Honor 400 and 400 Pro phones launching on May 22nd, 2025, making Honor the first smartphone maker to implement this particular Google AI technology.
• The tool transforms static images into five-second videos in either portrait or landscape orientation, taking one to two minutes to generate each clip.

How it works: The AI video generator is built directly into the Gallery app on the new Honor phones with a deliberately simplified interface that doesn’t allow for text prompts.

• Users simply select an image, and the AI interprets what kind of movement or animation would be appropriate without additional guidance.
• According to testing, the technology works best with straightforward subjects like people or pets, creating relatively realistic movement.

Mixed results: Early demonstrations show the technology produces unpredictable and sometimes bizarre outcomes depending on the source image.

• When presented with complex subjects like vintage cars or multiple people, the AI created physically impossible rotations and unrealistic scenarios—including a women’s soccer game with “27 players across three teams.”
• In one humorous example, when processing a Vincent Van Gogh self-portrait, the AI generated a video where a pigeon flies out of the painter’s eye.

Monetization strategy: Honor is offering limited free access to the feature before transitioning to a subscription model through Google.

• Honor 400 owners will receive free access for the first two months with a daily limit of ten video generations.
• While details of the eventual subscription cost remain undisclosed, the feature represents a new potential revenue stream for both Honor and Google.

Broader context: Google already charges for similar AI video generation capabilities through its enterprise and premium services.

• Image-to-video generation using Veo 2 is listed in Google Cloud at 50 cents per second of output, but is currently restricted to “approved users” only.
• Gemini Advanced subscribers can already access text-to-video generation using Veo 2, but not yet the image-to-video capability that Honor is now offering.