As the world of artificial intelligence for video continues to turn, many of those first AI apps and tools rushed out the door a few months ago are starting to get their second (and sometimes third or more) waves of updates. With the goal of refining bugs and smoothing out their machine-learned processes, this is usually how it goes for all new products and features.

So, it’s with no surprise that Runway ML, which we’ve covered before, has just released a major update to its image-to-video artificial intelligence technology, which they claim is going to significantly improve the quality as well as provide a new level of smoothness.

If you’re someone who’s messed around with these different AI video apps, you probably know the issues this new update is addressing as a type of ebbing, crawling effect has been noticeable in a lot of AI powered video clips.

Let’s look at this new update and explore how it looks, plus go over some insights into what this could mean for the future of AI for video as it continues to evolve.


A Look at Runway Gen-2

Launched back in the early days of AI (which basically means 2018 at this point), Runway ML has been one of the larger and more popular AI video editing tools for some time now. While racking up quite a bit of funding in investments, Runway has since released a mobile app as well as notably been used as an AI tool by the editors of Everything Everywhere All at Once.

With an intuitive app design and a free-to-sign up starter version, Runway signups have been through the roof. However, like many AI video apps and tools, it wouldn’t be surprising to learn that most who check out these apps are there to just mess around as the final results might not be ready for every person and project need.

Which leads us to this latest update — and obviously what Runway has been hard at work developing.

Image-to-Video AI

While this isn’t a new feature per se, Runway has announced that they’ve refined their existing AI technology for improved video generation. And, unlike other text-to-video features and apps you might see, this tech is based around an image input for true image-to-video AI generation.

You can see some examples of this improved image-to-video in these clips from Runway’s Twitter (aka X — I guess):

As you can see, these video clips — while still not perfect — are indeed looking better and more natural as opposed to the very wavy and odd previous versions. Still, it’s hard to look at these and not sense that something is just a bit off and unnatural.

Here are some more examples made by actual users:

What is the Value of AI Right Now?

Now, the question at the heart of this news — and pretty much everything else that has to do with generative AI video — has to be about the value it provides to creators and video editors. Have these tools truly risen to replace the video editing profession altogether? Probably not yet.

And while Runway does offer a free plan, it has a hard credit limit as well as limitations on video length as well as other upscale limits and watermarks. So if you’re at all serious about using this image-to-video tech you’re going to need to pay.

However, with plans ranging from $12 to $76 a month along with a credit system, the cost of use will need to be pretty tightly aligned with any value it provides. So the question really comes down to just how much you might want to use this AI for your own projects.

If you are interested though, you can find out more about Runway on its site and explore its pricing options here.