AI can't reject your no good, very bad idea

By Parmy Olson

Yael Biran has worked for the last 25 years as an animator for mostly corporate clients, capitalizing on her talent for colorful illustration, movement and figuring out what her customers want but don’t know how to articulate. Recently, she sat on her couch at home and was “freaking out” about her life’s work. She had several big expenses on the horizon, and her usual workflow of about a dozen annual projects had dwindled to three in the past year.

The reason was obvious: artificial intelligence.

More of the clients and creative agencies she worked with were trying to do animation work themselves, and she suspects they were using AI tools to do so. Biran is resigned to what that means for her. Another animation veteran she knows just retrained to be a gardener, and Biran is mulling new paths too, but she has a stark warning for what clients are about to lose: the people who challenge your terrible ideas.

“What we give to clients is the ability to say ‘no’ to their ideas,” says Biran. “They’re not visual people, and they know what they think they want. And then a lot times it really needs tweaking. Sometimes in a major way.”

The content that generative AI models can now conjure can sometimes look as good as anything created by humans. Creative agencies have been using tools from New York-based video generation startup Runway to develop concert backdrops for Madonna and graphics for CBS’s “Late Show With Stephen Colbert,” often saying it saves them hours or even weeks of work. Earlier this year, actor and movie producer Tyler Perry said he was halting a $800 million studio expansion because of the release of OpenAI’s video generator Sora, whose capabilities he called “mind-blowing.”

Critics of the tech say that will lead to a flood of boring, derivative work in film and TV since AI tends to spew a pastiche of pre-existing art, like Biran’s swirling watercolor figures or the quirky cartoons that she creates. But when companies use AI to generate animations for their own marketing, the effect could be worse thanks to the relative lack of visual, creative thinkers among their ranks. They’ll use AI tools to churn out graphics that — as with Hollywood’s overuse of CGI — look impressive but fail to make a meaningful impression on other humans.

One design agency, for instance, tried making a short animated film graphic for a British health-care provider that was meant to train doctors on their bedside manner. The script said medical professionals should listen carefully to their patients and avoid behaving like they were going through a checklist. Yet the resulting animation showed a physician sitting with a patient and a giant list being marked off above them. That’s not how visual communication works, says Biran. “People will see a checklist and go away thinking, ‘checklist.’”

When corporate clients try putting together a slideshow, they’ll also gravitate toward displaying some the same text already being spoken in a presentation, but that can make a presentation more confusing. There’s a reason for the phrase “A picture paints a thousand words.” Images can elevate subtext and advance a message, but figuring out which images are best requires people who are skilled at thinking visually, like Biran. “We think in pictures, and we gravitate towards metaphors,” she explains. “And so we can help identify the subtext.”

Corporate clients often believe they are visual thinkers too, perhaps because so much of the content people see online now is visual on platforms like TikTok, YouTube and Reels. But passively consuming graphical content doesn’t mean you can do a decent job making it.

“AI can clearly enhance our capabilities, but some clients are now questioning the need to hire creatives,” says Leila Makki, who runs a video production company for brands and agencies. A big reason may be the anticipation of OpenAI’s Sora. “They’re genuinely uncertain, but dismissing creatives for AI is shortsighted and counterproductive.”

Businesses would do well to avoid outsourcing too many aspects of creative work to artificial intelligence, even as they shift much of their marketing spending — which for North America and Europe tends to hover at around 9 per cent of capital expenditures — to generative AI. Biran predicts that in a few years, more companies will realize they need visual thinkers “and they will circle back and ask for our help.”

That may be an optimistic view, considering generative AI models are only becoming more sophisticated, with the possibility of greater reasoning capabilities to boot. But companies will also need people who understand visual communication to challenge their ideas, and they won’t get that from sycophantic AI models that don’t experience color and sound. They’ll get that from humans, who will need to get paid.