On Tuesday morning, the heads of Sony and Take-Two shared a stage in New York and agreed on the record that AI is not gonna replace human creativity. Newsflash folks, the grass is green, the sky is blue. We get it. I mean that without the snark, mostly. All right, a little bit of snark. If you make things for a living, you've been saying this since the first Midjourney image hit your feed, or you started building things in Discord with it. It lands differently coming from the people who sign the budgets.

So here's what they said. The Sony Group CEO, Hiroki Totoki, and Take-Two CEO Strauss Zelnick spoke at the Paley International Council Summit. Um, I read it through GameSpot, and IGN reported on it first, I think. Regardless, Totoki said AI can't replace humans in entertainment. Because, if you can believe it, creativity belongs to people. He also said he tells his own staff to use AI as long as rights holders, creators, and the artists stay protected, which is a pretty good policy.

Zelnick built his case a little bit differently. "AI is datasets, it's compute, and it's language models," he said. "Datasets look backward. Creativity looks forward." I can tell you this from my experience working at agencies over the years, as well as running creative departments as a creative director. It was interesting, he told a Home Alone story. He oversaw the release of Home Alone at Fox. Nobody expected a hit, and he says it's gonna be the highest-grossing comedy ever. His conclusion: datasets aren't gonna get you there.

He also passed along a line from one of his division heads. "You're not gonna be replaced by AI," he said, "but you might be replaced by someone who knows how to use it." That's the sentence in this blog post that, like, you should screenshot.

This isn't a new thing. Zelnick made the same argument on a Paley stage a whole year ago. PC Gamer's report from the 2025 summit has him calling AI a "parlor trick" for a reason. It only predicts from past data. Of course, then, the November earnings call comes where he said Take-Two isn't using AI as an excuse to cut headcount. In March, he told The Game Business that AI tools "may help you create assets, but that won't help you create hits." His point was that thousands of mobile games launch every year and only a handful of them become hits. What it comes down to is that assets are cheap, hits are very rare. Anyone telling you a prompt replaces a creative director is full of shit, and hasn't spent enough time in the world to know better.

 

Landscape collage with paper clouds and green field cutouts.

Gamers in particular are very fickle. I'm one of them. Everyone will jump on board a new release, fall in love with it, play the shit out of it, and then dump it for the next shiny thing that comes along because they've either been grinding too much or it's lost its luster in any other number of ways. Take it from me, as a gamer with my group of gamer friends... we're constantly in a spin cycle of looking for the "next thing" to play.

Shuhei Yoshida, the longtime PlayStation executive, made a similar point to this in April of 2023 when he was still at Sony. He told The Guardian that AI is a tool and somebody has to use it. PC Gamer's write-up has him saying, "Creativity is more important, the direction, how you envision what you want." He's no skeptic, and he's retired from Sony now. By late 2025, he was telling Game*Spark that developers from indie to major studios use AI in all aspects of game development. And he sees agents built into games down the line, which could be really cool or could be a complete fucking nightmare. His line was never "AI can't do it." It was "AI needs a person with direction." Three and a half years later, that's the part that hasn't moved.

Neither company is sitting this out. Don't read this post as "AI bad." Sony's own numbers show that more than 50,000 employees across 210 teams using its internal LLM, with over 300 AI projects tested and more than 50 in regular use. Take-Two is running hundreds of pilots while saying gen AI has no part in GTA VI. The real position in this seems to be narrower. Tools make assets. People, especially creative people, make the calls. It's the same line that I drew in the "You never had a choice, run with it" post that I wrote, and it's nice to see it from the other side of the table.

Personal experience as a creative director all these years really can help me understand this argument. I've seen time and time again that people, not datasets, form way better ideas. This is especially important with visual design. Designers use context. Designers use experience, creative juice, late nights, and datasets to come up with ideas. I've always said that design is different than traditional visual art in a very special way. Art can very simply be for you and no one else. People may love it, hate it, or never see it... and that's okay. You really don't have to give a shit if people judge your art.

Design is a little different because it's art that solves a problem, and sometimes those problems can be very complex. Many times, those problems are directly connected to datasets. The datasets will inform the direction, but they sure as fuck don't do the work for you, and AI doesn't seem to do a good job of being able to figure it out right now either. It's not just about the fact that relying solely on AI will remove the magic of five designers standing around a table with all of the rough designs and thoughts on the table, it's the fact that AI will not be able to effectively replace the five designers standing around a table.

So, what do you do with this? When a client is asking you why you or I cost more than a prompt, now you've got two CEOs on record. Use the tools for grunt work. Let the tools rename layers, remove backgrounds, draft that super dull first pass. Keep the concept, the taste, the craft, and the final call to the adults (the humans). Zelnick's division head is right that someone who knows the tools will definitely beat someone who doesn't. He's also right, I think, that knowing the tools isn't the same as having the actual idea.

I typically use AI for inspiration, palette development, ideation. It's my... dumb/genius futuristic assistant.

All in all when I see news articles like "Important Person States that AI isn't as Good as Real Creative People", I roll my eyes to a degree, because I've stuck to the same feeling and statement for years.

No shit. Grass is green, sky is blue - humans make the hits.