Google launched on Tuesday, April 15, I see 2, its latest artificial intelligence model (AI) to generate videos of eight seconds in high resolution, aimed at Gemini Advanced subscribers worldwide. It will be available in all languages compatible with Gemini, such as Spanish.
With this tool, an eight -second video clip can be generated with a resolution of 720p, which the user can download as an MP4 file in horizontal format 16: 9, as indicated by the company in a statement.
I see 2 has a monthly limit in the amount of videos that can be generated. However, the technological giant did not indicate in its statement the exact amount; He only pointed out that when the user approaches the deadline, he will be notified.
I see 2 represents a great advance in the generation of video, designed to produce detailed high -resolution videos with cinematographic realism, the company notes.
Google also offers I see 2 through whisk, which allows you to create new images using text indications and images, and now encourage them to turn them into videos.
Although a brand of water will not be seen at first glance indicating that the video was generated by AI, Google states that all the videos generated with I see 2 are marked with Synthid, a digital water mark embedded in each frame, which identifies that the videos were created by artificial intelligence.
The company also points out that it has an exhaustive evaluation process to prevent the generation of content that violates its policies.
Meanwhile, if a user asks to reproduce the image of a real person, receive the following message: I can’t generate that video. Try to describe another idea, according to EFE.
Google faced controversies with its first generators of photorealist images, after inaccurate and biased responses.
The first models received criticism for having generated images of an Asian woman or a black man by requesting an illustration of a 1943 German soldier, a date on which Nazism was booming in that country and racial diversity was practically non -existent.
Today, EFE asked Vie 2 to recreate a scene from World War II, and the model generated a video with three white men with helmets, uniform and rifles.
Among the examples shown today by the company, there was an animated video of a man eating an ice cream in the style of the Japanese animation studio Studio Ghibli.
This is no accident. At the end of last month, social networks were flooded with portraits and memes generated by the latest model of OpenAi, Chatgpt-4o, inspired by the Japanese anime style of films such as Chihiro’s trip, my neighbor Totoro and El Niño and La Garza.
I see 2 will now compete with Sora, from OpenAi.