Alphabet Inc.’s Google is rolling out a new AI-powered video generation feature for subscribers of its premium Gemini plans, in a move to stay competitive in the rapidly evolving generative video space.

Starting Thursday, users on the Google AI Ultra and Pro plans in select regions will gain access to a tool that transforms photos into 8-second video clips using the web version of Gemini, with mobile access set to follow later this week.

The feature, which creates 720p MP4 videos in landscape format, combines user-uploaded images with text prompts to generate clips complete with sound. Google says the tool is based on Veo 3, its latest video generation model unveiled in May and previously only available via its standalone Flow tool.

The update brings the video-generation capability into Gemini’s core chat interface, directly challenging rivals like OpenAI and video-focused startup Runway AI Inc.

It also pits Google against fast-growing players in China, including Alibaba, Manus.AI, and Kuaishou, all of which have recently released their own AI video tools.

To address safety and ethical concerns, Google has implemented guardrails to prevent the generation of videos using images of celebrities or public figures, and bans outputs that promote harmful or violent content.

However, early tests reveal challenges.

Bloomberg News reported instances where the tool altered facial features or even the race of subjects when asked to animate personal photos, raising questions about reliability and bias in the system’s output.

The launch is part of Google’s broader push to integrate more advanced generative tools into its ecosystem, as it races to maintain relevance in a market defined by innovation and rising user expectations.