Adobe headquarters in San Jose, California. (Photo by David Paul Morris / Bloomberg)
Adobe Inc. unveiled artificial intelligence tools that can create and modify videos, joining Big Tech companies and startups in trying to capitalize on demand for the emerging technology.
One feature, integrated into Adobe’s video-editing software, Premiere, will let users extend video clips using generative AI, the company announced Monday at its annual product conference in Miami. Other tools, available online, will let users produce video from text prompts and existing images.
While OpenAI, Meta Platforms Inc. and Alphabet Inc.’s Google have shown off AI video generators, Adobe is the first big software company to have it widely available for customers. Some startups, such as Runway AI, have already released their video-generating products publicly.
"What we hear when we talk to our customers is it’s all really cool but they can’t use it,” Ely Greenfield, Adobe chief technology officer for digital media, said of the competitor’s technology.
Customers want AI features within applications they already use, Greenfield said. Adobe’s new video models are "designed for real work flows and integration into tools,” he said.
Over the past year, Adobe has focused on adding generative AI features to its portfolio of software for creative professionals, including flagship products Photoshop and Illustrator. The company has released tools that use text to produce images and illustrations that have been used billions of times so far.
Adobe has sought to differentiate its models as "commercially safe” due to cautious training data and restrictive moderation. For example, there are certain faces Adobe will block if users try to generate videos of them, Greenfield said.
Rivals have come under fire for widely scraping the internet to build AI models.
Adobe’s video models were trained primarily on videos and photos from its vast library of stock media for marketers and creative agencies, Greenfield said. In some cases the San Jose, California-based company used public domain or licensed data, he added. Adobe has offered to procure videos for about $3 per minute from its network of creative professionals.
OpenAI’s demonstration of its video-generation model Sora earlier this year ignited investor fears that Adobe could be disrupted by the new technology. The company’s shares declined 17% this year through Friday’s close.
Adobe isn’t yet charging for the use of its AI features beyond its standard subscription fees. Each user is allotted a number of credits for AI generations, but the limits aren’t being enforced for most plans, Greenfield said. In the future, Adobe may charge more to use its video-focused AI than its similar tool for photos, company executives have said.
At its conference, Adobe also announced improvements to other software, such as making it easier to view 3D content in Photoshop. The company also is working on developing AI models that can generate 3D graphics.
