In a rare OpenAI interview Chad Nelson faces the 'AI Slop' backlash and says "rising above the noise takes talent"

The conversation around generative AI art and creativity is getting louder, and messier. From visual tools like Sora and DALL·E to the rise of 'AI slop' criticisms flooding social media, the debate isn’t just about what AI can do, but whether it should be doing any of it at all

I caught up with Chad Nelson, Creative Director at OpenAI, while at Upscale Conference in San Francisco, to talk about whether AI can do art, why true creativity still needs a human eye, and what the next Spielberg might look like in a world where anyone can prompt a sci-fi epic in seconds… or can they?

Chad argues that AI can and will reach a point where anything can be generated and created, from vibe coding a game to creating an animated movie, but a breakout film, game or artwork will be harder to achieve as millions of AI-created films hit our screens.

The AI problem coming next

AI tools are democratising creation, or at least that's the pitch, but Chad sees a problem on the horizon when it comes to discovery. The new Spielberg could be an AI creative, but they could also be buried under a deluge of Ghibli memes and cat portraits.

"Discovery and finding eyeballs is going to be one of the great challenges," Chad says. "We have word of mouth, we have algorithms, maybe social media helps, but with the sheer quantity of content coming out, even within your niche, how do you get seen?"

The reality, Chad points out, is that even great content can drown in the 'slop' flood unless someone pays to promote it. "If there’s one thing Hollywood and record labels know how to do, it’s blanket the planet with promotion. Everyone knows when the next Mission: Impossible drops, whether they go see it or not."

AI is often touted as the great equaliser, but Chad's not ready to claim we’ll see a surge of AI-enabled Spielbergs just yet, and so that pitch to break down barriers feels a little hollow. But looking deeper, it means a level of artistry and knowledge will be needed to ensure anything AI-generated has value.

"It’s a lot of work to make a 90-minute film with real emotional impact. It’s not just one person – it’s wardrobe, production design, music, editing, direction, cinematography. There are very few people who are brilliant across all those."

Instead, Chad expects AI to empower a wave of simpler, more focused indie-style stories. Big, emotionally rich action scenes? Still hard. "Even now, people complain that action scenes feel empty. AI or not, making one with real dramatic weight is still an art."

I ask Chad about the term 'AI slop', often used online to describe lazy or low-effort AI content, and he doesn’t dodge the five-legged elephant in the room.

"It goes back to curation," he explains. "I can ask for 1,000 versions of a woman in a 1950s dress playing chess, but it takes a creative eye to find the Queen’s Gambit composition in that pile."

In other words, generating images and video is easy, and getting easier, but choosing what’s good, and understanding why it’s good, is where the real skill lies. That, Chad argues, is the future of creativity: using AI as a collaborator, not a replacement.

"The one brilliant thing about ChatGPT is that it can give you an answer. The one not-so-brilliant thing? It doesn’t know if that answer’s right,” he jokes.

Even as AI models improve, Chad believes standing out will become harder, not easier. "Anyone can make something that looks decent. But rising above the noise? That’ll take talent."

The rise of remix artists

Nelson doesn’t pretend current systems are capable of pure, groundbreaking innovation. Instead, he sees them as 'remix artists', capable of "creating combinations of imagined things in unique ways".

"There are very few humans who are truly original innovators,” he says. "We all see the same visual cues, we wake up in the same world. So it’s fair to say a lot of what we create is derivative, even if we don’t want to admit it."

But could AI ever learn to be truly original? Nelson’s not sure, but he’s interested in the question.

"What if we taught it how to use a charcoal stick or a wooden brush without giving it any data about the real world? Would it create anything meaningful, or just abstraction? That would be fascinating to see."

Who is AI actually for?

That last comment offered a glimpse into the world of OpenAI, where seeing what could happen with AI is as fascinating as why or how it exists. We often fixate on a creative world divided between professional creators and casual artists.

"Are we building it for Spielberg, or for my mom who just wants to make something fun? Or are we trying to do both?”

For OpenAI, the answer seems to be… everyone. But the real breakthroughs, Chad believes, will come from third parties building specialised versions of these tools, ones that are tailored for fashion designers, filmmakers, artists and others.

Ultimately, Chad's says he's worried about AI replacing artists. He’s more focused on augmenting them, and keeping the soul in creative work. It's often easy to view the AI art debate as for and against, but it's more nuanced.

"I’ve heard people say a third of their company is all in on AI, a third is resisting with pitchforks, and a third’s still figuring it out," explains Chad. "I can’t fix the bottom third, but I can try to show the middle what’s possible – and what’s not."

Creativity, in Chad's view, isn’t about using the newest tool or the latest AI model. It’s about knowing what to say, to analyse what works and why – and having the taste, training and talent to say it well, to self edit and find something original and artistic in the output.

“AI gives you a thousand options," reflects Chad. "But only you know which one is worth choosing."