Ticker

6/recent/ticker-posts

Google Veo 3: AI that generates ultra-realistic videos is taking the web by storm

Google Veo 3: AI that generates ultra-realistic videos is taking the web by storm

Last week, at Google I/O, Google kept making a lot of announcements about artificial intelligence. And while some announcements about the evolution of its search engine may have caused serious concern among news site publishers, today it's its new video generator that's causing a lot of buzz.

Among the tools that attracted a lot of attention at this event, the company unveiled Google Veo 3, a new solution for generating videos using artificial intelligence... with sound!

A complete video with a simple prompt...

In the Google Veo 3 demonstration videos that have recently circulated on the web, we can see that using this tool is very intuitive. From a simple text prompt, users can generate a video with sound, dialogue, or even sound effects.

From street interviews, to movie trailers, to fake news reports, the possibilities offered by Google Veo 3 represent a major advance in artificial intelligence…

A saving of $500,000 thanks to Google Veo 3

While the images produced by Google Veo 3, many of which were revealed on the social network X, are very impressive in their realism, it is in another aspect that Google has surprised its audience…

Until now, if video generation tools like OpenAI's Sora could produce visually successful results, this was limited only to this part.

With Google Veo 3, users can now generate videos with sound, including very realistic dialogue. Among the many videos that have been shared online, it is the advertising video by PJ Ace, a director specializing in AI, that has generated a lot of interest.

According to the creator of this pharmaceutical advertisement, the production of this video with real actors and a film crew would have cost at least $500,000. Enough to seriously worry the audiovisual world, beyond the problems posed by the disinformation facilitated by Google Veo 3…

A tool that is still very limited

While Google Veo 3 is astonishingly realistic, this information should be balanced with a few remarks… Indeed, video sequences are currently limited to 8 seconds, which makes it possible to very quickly identify videos generated by Google Veo 3, in addition to complicating the production of videos in longer formats.

It should also be noted that the video generator is currently only available in the United States, although it is probably only a matter of time before it is accessible on the European continent.

Finally, Google Veo 3 is currently available to users who have opted for an AI Pro subscription at €250 per month. While this should not prevent some disinformation videos from appearing on the web, these constraints should limit its use for the time being...

Post a Comment

0 Comments