Be part of our every day and weekly newsletters for the most recent updates and unique content material on industry-leading AI protection. Study Extra
Regardless of going through controversy — as many AI mannequin suppliers have these days — over its knowledge scraping and coaching practices, the New York Metropolis-based startup Runway is plowing forward with new marquee options for its reasonable generative AI video platform.
At the moment, the corporate introduced on its X account and its Discord server that Gen-3 Alpha, the brand new AI video mannequin it unveiled again in mid June 2024 able to producing extremely reasonable video in seconds from easy textual content prompts, now accepts nonetheless photos as prompts as properly.
The consumer merely has to navigate to Runway’s web site, click on on the “try Gen 3-Alpha” language and they are going to be taken to a display the place they will add imagery and/or proceed to enter textual content prompts to information new AI video generations which are 5 or 10 seconds in size (the consumer chooses). A ten-second generated video requires 40 credit by way of Runway’s pay-to-play and subscription tiers, whereas a 5-second technology prices 20 credit.
Checks reveal Runway’s Gen-3 Alpha image-to-video replace is surprisingly quick whereas retaining top quality
We examined the function briefly on my private Runway account and I used to be personally blown away by each the pace (it took lower than a minute to generate the video from a nonetheless picture I supplied) and the standard. See my nonetheless (generated with separate AI service Midjourney) and video made with Runway Gen-3 Alpha embedded under.
Security measures in place
The mannequin mechanically makes an attempt to detect and block the creation of video from express nonetheless imagery or well-known figures resembling politicians.
The function additionally made its approach on-line early briefly for some Runway subscribers over the weekend earlier than being yanked down after which put again up at the moment.
Preliminary examples abound
On its X account, Runway additionally highlighted 10 fascinating movies generated from nonetheless photos with the Gen-3 Alpha mannequin.
As Runway posted in its tweet: “Image to Video is major update that greatly improves the artistic control and consistency of your generations.”
Runway’s co-founder and CEO Cristóbal Valenzuela launched the brand new function on his personal private account on X with the easy phrase “it’s time” adopted by a submit of his personal image-to-video creation.
The video AI race is neck and neck — pending litigation
Runway is certainly one of a number of firms — together with OpenAI with its Sora mannequin, Kuaishou Know-how with its Kling AI mannequin, Luma AI with its Dream Machine, and Pika with its self-titled mannequin — vying to supply quick and top quality generative video to customers.
Nevertheless, OpenAI’s Sora nonetheless stays personal for now, whereas the opposite fashions are accessible to members of the general public to make use of.
However with Runway’s new nonetheless image-to-video replace promising one more approach for customers to make Hollywood-quality movies on the fly, it’s clear that the whole area of filmmaking and video creation is being upended like by no means earlier than.
However will the courts stifle the innovation on the behest of aggrieved artists and rights’ holders of copyrighted materials?
Runway is, in any case, amongst a number of AI firms at the moment being sued in school motion lawsuits by creators who allege that the final observe of AI firms scraping and coaching on publicly posted materials — together with copyrighted materials — with out categorical permission, authorization, compensation, or consent is a violation of copyright legislation. How the courts rule on this challenge will little question vastly influence the current and way forward for AI video and artistic instruments.