The video reviews Google’s AI video generator VO3, highlighting its impressive ability to create realistic scenes with people talking and gameplay content, while also noting its limitations in full-body animations, vertical videos, and accurate text or lip-sync. Despite these restrictions, VO3 is considered the most advanced AI video tool available, offering significant potential within its current capabilities.
The video provides an in-depth review of Google’s latest AI video generator, VO3, which is accessible through the Google Flow platform and Gemini. The creator begins by demonstrating how to start a project and input prompts to generate videos, highlighting the tool’s impressive ability to understand detailed scene descriptions and produce realistic videos, such as Zoom calls with synchronized lip movements and dialogue. Despite its strengths, the reviewer notes some flaws, like gibberish or misspelled text in videos, and emphasizes that VO3 excels at creating scenes with people talking but struggles with generating full-body videos and accurate text on signs or billboards.
The reviewer explores various use cases, including generating gameplay videos of popular titles like GTA 6 and Starcraft, as well as anime scenes and music performances in different languages. While VO3 can handle singing and some language variations, it faces limitations with full-body animations, dynamic dance movements, and complex scenes involving multiple camera angles or intricate actions. The tool also allows control over dialogue tone and emotion, making it versatile for creating different atmospheres, but it still cannot produce accurate accents or lip-sync existing voices, which limits its ability to generate consistent characters or celebrity impersonations.
Further testing reveals significant restrictions, such as the inability to generate vertical videos suitable for TikTok or Shorts, and challenges with image-to-video conversion, especially when trying to create talking characters from static images. The tool’s censorship policies also prevent generating images or videos of specific celebrities or existing people, although some workarounds like uploading images and bypassing filters are possible. Additionally, the platform cannot control character voices or lip-sync pre-recorded audio, which hampers efforts to produce consistent, realistic characters across multiple scenes or videos.
The review also highlights the limitations in generating sports scenes, such as soccer or tennis matches, where the quality and realism are poor, with frequent errors in gameplay and action depiction. The creator discusses the pricing plans, noting that Google offers a pro tier at $20/month with limited credits and an ultra tier at $250/month for more extensive use. Currently, new users can access a free trial or a 15-month free student plan, allowing limited but valuable access to VO3. Despite its imperfections, the reviewer emphasizes that VO3 remains the most advanced AI video generator available, outperforming other competitors significantly.
In conclusion, while VO3 demonstrates remarkable capabilities in creating scenes with people talking, gameplay, and animated content, it is still far from perfect. Its restrictions on full-body videos, vertical formats, celebrity likenesses, and voice control limit its application scope. Nonetheless, it stands out as the leading AI video generator, offering impressive results within its current limitations. The creator encourages viewers to share their experiences and thoughts on VO3, and promotes their newsletter for ongoing updates in the rapidly evolving AI landscape.