It may well be that we have the first AI capable of generating videos from text. Gen-2, a tool developed by Runway, works much the same as Midjourney and its equivalents. If for now, the technology is still in its infancy, the results obtained are already very impressive and promise a real revolution in the field of video.
If things particularly accelerated at the beginning of 2023 with the invasion of ChatGPT and its descendants, it has been possible for some time now to generate images by artificial intelligence. This novelty has also been the source of many controversies, particularly around the question of copyright and the protection of artists. However, this does not prevent the tech giants from looking very seriously at the question, as Microsoft recently showed us with its own tool integrated into Bing.
Today, therefore, the generation of images by AI is already at a relatively advanced stage. The same cannot be said, however, of video, a fantasy necessarily much more complex to achieve with its share of constraints. The start-up Runway has nevertheless devoted itself to meeting the challenge, which is moreover with some success. Indeed, the latter could well be at the origin ofa (not so) small revolution in the video industry.
Discover Gen-2, the video equivalent of Midjourney
This revolution has a name: Gen-2. This AI arrives, as its name suggests, after Gen-1, a tool launched in February already capable of retouching videos and applying filters to them from a text command. Gen-2 goes much further, since it is capable of fully generate a video without a prior templatesimply from a user description.
In its Modes 02 and 03, Gen-2 even offersadd an image to its description to stick as well as possible to the expected result, or even to dispense completely with text and just rely on the picture. The other Modes, meanwhile, continue the work of Gen-1 by offering various video editing options.
Related: Netflix Replaces Human Artists With AI To Generate Cartoon Sets
Admittedly, for the moment the generated videos are not yet perfect. It is clearly visible that these have been generated by AI, which has difficulty in developing realistic images in large quantities. Still, getting to this level in just a few months is truly amazing.t. There is no doubt that Gen-2 will trigger profound changes within the industry.
Source : Runway
