OpenAI’s Sora tool reveals industry divide over AI and creativity

Sora, OpenAI’s new text-to-video platform, is being hailed as a technical triumph within the advertising community. But the industry remains divided as to whether the platform will augment creativity or muzzle it. As with other generative AI tools, Sora begins with a text prompt, from which it can produce an original video based on training data. The difference, however, is in the quality of this output, which is far more advanced than counterparts such as Runway 2, experts told Ad Age. Sora’s videos are of photorealistic clarity, reflecting a persistence of detail and style as well as an accurate comprehension of the physical laws of motion. These developments, along with its ability to generate a video up to one minute long, are why the new tool is already affecting strategies for the future of production. “It’s unbelievable for creativity … I’m having a creative anxiety attack,” said Dan Gardner, co-founder and executive chairman of agency Code and Theory. Sora is currently in beta and only open to a select group of testers; its public release date is unclear. When OpenAI revealed the first videos of Sora just over a week ago, Phil Russell, a video editor at production agency The Production Factory, also felt bowled over. But 20 minutes later, that feeling became something more like hollowness as he realized that Sora could accelerate demand for unimaginative and stock-like ad creative. “We’re just filling more empty content, as opposed to [bridging] any real gap in the industry,” he said. Sign up for Ad Age newsletters From influencer marketing to agencies, get the latest news and analysis delivered to your inbox Click here The divide between Gardner’s optimism and Russell’s exasperation reflects a deeper rift within the ad community over the role that AI should play in the creative process. On one side, marketers—many of whom are tech-forward and focused on innovation—see AI tools like Sora as a boon to the industry and a multiplier for creativity. On the other side, practitioners—many of whom work within the creative process—feel frustrated by what they view as a compromise of artistry for the fast and cheap tricks of a trendy technology. “We’re talking about taking significant cost and labor out of video production,” said Jay Pattisall, VP and principal analyst at research company Forrester. “I can understand why there’s lament.” Sora as an inflection point Previous generative AI tools have stirred excitement at launch, but the particular advancements of Sora make it stand out, said Kaare Wesnaes, head of innovation for Ogilvy North America. The model seems to have a much better understanding of physical laws, portraying motion in a persistent and mostly realistic way. Other text-to-video platforms, such as Runway 2, have yet to demonstrate the same level of comprehension, said Wesnaes. Meta’s Emu platform and Google’s Lumiere platform are even more so hampered by glitchy movements. Maximum video length of these other tools—anywhere between three and 18 seconds—is also a fraction of that of Sora. With the ability to create videos of up to one minute long, Sora can allow creatives to not just explore an idea, but more fully tell a story, according to Elav Horwitz, head of applied innovation at McCann Worldgroup. “With Runway, I can’t tell a story,” she said. Most of all, Horwitz added, people seem to be impressed by Sora’s photorealistic clarity. Numerous videos have already gone viral for making the absurd appear lifelike, such as an instance depicting two dogs hosting a podcast at the top of a mountain, which garnered over 7 million views on Twitter. https://t.co/uCuhUPv51N pic.twitter.com/nej4TIwgaP— Sam Altman (@sama) February 15, 2024 But Sora’s unprecedented quality is also why it feels like an inflection point in the greater AI conversation, according to Russell. No longer does the output, upon first inspection, appear AI-generated, he said. “This content will pass as usable to the average person, and when marketers start to allow that to happen is [when] we’ll have a problem,” he said. Whereas Russell is worried about the diminishment of artistry in advertising, Gavin Knight, a freelance writer and director, is more concerned that professional opportunities will be reduced. Knight has done commercial shoots for brands all over the world, including in Tokyo and Lagos, Nigeria. Among the preliminary videos released by OpenAI showcasing Sora’s abilities were shots of Tokyo and Lagos. “I can’t tell the difference between what I shot there and what Sora is generating,” he said. Knight wonders if, soon enough, clients will use Sora over hiring him to produce footage of a foreign country and splice their product with the video. His workload has already shrunk following the pandemic and spillover from the U.S. writer’s strike last year. AI, he fears, could make for a third act of reduction. Forrester’s Pattisall said Sora is “close to production quality,” and that its ability to fit the 15-, 30- and 60-second formats of most video players is a reason that advertisers will be even more intrigued. Others, like Wesnaes, do not view Sora as yet producing the kind of video that should warrant much concern. Wesnaes could see the AI platform creating B-roll, as well as stock material for brands, but nothing of “Super Bowl quality,” he said. OpenAI agrees, and readily admits on Sora’s landing page shortcomings that still need work. Complex physics, for example, may not be accurately reflected in videos. And ambiguities in a prompt can still easily confuse the platform. Common ground for a path forward With so much investment fueling AI development, creative practitioners have good reason to fear that their voices are at risk of being drowned out. But to reject the technology would be unwise, said Andrew Shepherd, a freelance videographer. “We can use these AI tools as just tools,” he said. Forrester’s Pattisall sees a similar path forward for the ad industry at large, in which AI can be taken for both its exciting implications and concerning ones, not either/or. Through this approach, creatives must embrace developments in technology, while understanding that the need for creativity is not going away. Indeed, originality was found to be the most significant factor in lowering a marketing job’s likelihood to be automated by AI, according to a report conducted last year by Pattisall and fellow Forrester analyst Michael O’Grady. The degree to which Sora will actually shake up advertising’s creative process will depend on how widely it is used once it becomes publicly available, Pattisall said. Upon launch, both creatives and tech leaders will need to apprehend a series of other obstacles introduced by AI. Deep fakes could become a much bigger issue as a result of Sora’s realism, said McCann’s Horwitz. Brands have been victims of this tactic, and are likely to be affected by forthcoming federal policy addressing deep fakes. Ogilvy’s Wesnaes expects Sora will eventually drive more AI-generated ads, and not the kind that are purposefully clumsy, as seen by numerous marketers over the past 15 months. But these efforts could also attract a higher level of scrutiny from consumers used to the execution of traditional advertising. “Now that you can create something of proper length and quality, you can create something that could receive pushback,” said Wesnaes. AI marketing blog How brands are using ChatGPT, DALL-E 2 and other tools Read more here

OpenAI’s Sora tool reveals industry divide over AI and creativity

Featured Stories