AI-generated video has gotten good. Should designers try it? – Business of Home
You’ve got ChatGPT writing copy for social media. You’re using Nano Banana to tweak renderings. You’re feeding your financials to Perplexity. But just when you thought you had mastered all the shiny AI tricks, our tech overlords have given the robots a new one: video. Buckle up.
To be clear, AI-generated video is not really new. But up until this year, it was either expensive, clunky or low-quality. Over the past six months, tech giants have rolled out a variety of new tools that make AI video fast, easy, startlingly good, and cheap (or free).
As so often is the case, it was OpenAI (the maker of ChatGPT) that won the biggest headlines, with its late-September release of Sora 2, a new model and accompanying app (just called Sora) that allows users to create 15-to-25-second videos with simple prompts. The tool has inspired awe and controversy in equal measure: At first, it took a laissez-faire attitude toward copyright law, allowing users to create videos of Iron Man slipping on a banana peel or Pikachu getting arrested for a DUI. It also drew concern over potentially harmful deepfakes with a feature called “Cameos” that allows users to insert their own likeness into AI-generated content. For example, you can make a video of yourself slipping on a banana peel, and, if you choose to share your Cameo, others can make the same (or worse).
But amid all that, there was no doubt that Sora had the goods. A month in, it has hardly become an everyday tool for designers, but tech early adopters are certainly taking notice.
“We had a lot of different AI models that were available,” says Amir Hossein Noori, an architectural designer and the founder of AI Hub, a company that provides artificial intelligence education and training. “Sora was this wow moment where we saw something I didn’t think was going to be possible this early on.”
Sora is free to use, but thus far requires an invite code. It’s not, however, the only game in town. Google’s Veo 3.1, another powerful video engine, is free to experiment with in Google Flow if you set up an account; Midjourney and Kling cost a little, but are quite good as well. There are slight differences in capability and interface among the four tools, but all of them can produce intriguing results without much expertise required. What good might they be for designers?
To Noori, the low-hanging fruit is design presentations. All of these tools allow the user to upload a still image and convert it into a video—you could feed Sora a rendering of a living room, for instance, and command it to “create a walkthrough of this space,” or “generate a video of a family enjoying this space.” (If your client has uploaded their own Cameo to Sora, you could virtually put them into the space.) Results will vary, and sometimes you’ll have to go through several rounds of refinements to get something good. But especially as the technology improves, it’s not a stretch to imagine AI video becoming a powerful presentation tool.
On the flip side, designers could also use AI-generated video to workshop their own concepts. Imagine sketching out a kitchen, then prompting an AI tool to show it in use. “Designers can explore the experience of a user themselves before finalizing the space,” says Noori. “I think it’s an aha moment for the interior designer to understand their own project before they show it to the client.”
Another, slightly lower-stakes use for AI video? “I think it’s helpful for social media,” says Jenna Gaidusek, a Charleston-based designer who coaches fellow designers on how to use AI. The technology, she says, can be a lifesaver for designers who are struggling to generate the kind of content—vertical video—now favored by social media algorithms: “You can take what you would typically do as a static post and turn it into a Reel easily, and it’ll reach more people on Instagram.”
Gaidusek, a restless experimenter, has been playing with a method to generate video before-and-afters, and she linked me to an artist and designer who was using AI to animate her paintings on social media. This is still very much the early, playful stage of the technology—I spent more time than I’d care to admit creating fairly ridiculous content, like a video of a raging Halloween party in Nancy Lancaster’s famous and extremely proper Yellow Room.
For now, AI-generated video sometimes feels a bit like a toy. But as it improves and adoption grows, designers will likely find cause to make it a tool.
