Google going head-to-head OpenAIWith the newest version of 's Sora video generation model, Veo 2, it says creates more realistic videos.

The company has also updated its imaging model Imagen 3, producing rich, detailed images.

Google said Veo 2 "better understands real-world physics and the nuances of human movement and expression." It is available Google Labs' VideoFX platform - but only on a waiting list. Users must register through a Google form and wait for Google to grant temporary permission at any time.

"Veo 2 also understands the language of cinema: ask it for a genre, point to a lens, suggest cinematic effects, and Veo 2 delivers in resolutions up to 4K," says Google's blog.

Although Veo 2 is only available to select users, the original Veo remains available At Vertex AI. Videos created with Veo 2 include Google's SynthID metadata watermark to identify them as AI-generated.

Google, however, admits that the Veo 2 may still hallucinate extra fingers, but it promises that the new model will be less hallucinatory.

Veo 2 competes against OpenAI's recently released Sora video generation model to attract filmmakers and content creators. Sora was in preview for a while before OpenAI made it available to paying subscribers.

Surprisingly, Google says that Veo was selected by human raters based on internal tests that determined "overall preference" (ie, which videos resonated better with audiences) and "adherence" (how well the videos matched the instructions given by the human creator). Sora and other competing AI models.

Google announced I see May A video shot at this year's Google I/O developer conference with actor-musician Donald Glover, aka Childish Gambino.

AI video generation still needs some work

AI video generation has long been an area of ​​generative AI where large model developers like Google and OpenAI regularly compete with and outpace relatively smaller companies.

RunwayMLOne of the pioneers of the AI ​​video generation, recently launched its advanced management Gen-3 Alpha Turbo model. Pika Labs Pika released 2.0gives users more control and allows them to add their own characters to the video. Forward AI announced a partnership with AWS for Take his models to Bedrock for enterprise use. Luma too His Dream Machine expanded generation model.

However, AI video generation still needs to convince both creators and viewers. After Sora's long-awaited release, people remained in disbelief about his possibilities as he continues to create figures against physics and anatomy. Users felt that it produced conflicting results.

More trailer from The Game Awards soon showed the distrust of the people what they perceive as "AI bias".

Some filmmakers have begun to do so accepting opportunities AI video generators can provide. Famous director James Cameron joined the board of directors Sustainability AIactor Andy Serkis has announced that he is building an AI-focused production company.

Google says it's seeing a lot of user interest. According to the company, YouTube creators have been using VideoFX to create backgrounds for YouTube Shorts to save time.

Updates for Imagen 3

Google recently updated its Imagen 3 image model, which it made available through its online Gemini chatbot, to offer more realistic and brighter images.

Imagen 3 can now accurately represent more art styles "from photorealism to impressionism, from abstract to anime." Google said the model follows the instructions exactly.

People can access Imagen 3 through ImageFX.

Daily insight into business use cases with VB Daily

If you want to impress your boss, VB Daily has got you covered. We give you the inside scoop on what companies are doing with generative AI, from regulatory changes to practical deployments, so you can share insights for maximum ROI.

Read ours Privacy Policy

Thank you for subscribing. See more VB news is here.

An error occurred.