Perhaps the biggest AI news this week was about what someone wasn’t doing with AI. James Cuda, the normally camera-shy CEO of Procreate, made a video where he came out and declared his app would not be adding generative features. Ever. He also shared his opinion of generative AI generally, and it’s fair to say he’s not a fan.
If you’re unaware, Procreate is an alternative to Adobe Creative Suite that’s popular with creatives. And the company is all-in on Cuda’s anti-AI stance. “We’re never going there,” the company’s Twitter feed declared about generative AI, linking to a curt web page that reinforced that point of view.
“I really f**king hate generative AI,” Cuda said. “I don’t like what’s happening in the industry, and I don’t like what it’s doing to artists. We’re not going to be introducing any generative AI into our products.”
Predictably, Cuda’s comments earned him instant hero status among the creative set, and they come at seemingly the perfect time: The backlash against AI is gathering lots of momentum, with many questioning whether the massive investment in the industry is justified. With regard to creative software specifically, there’s been ongoing skepticism of Adobe’s dips into generative AI, which culminated in an outcry over the company’s updated terms of service earlier this year (the company spent a lot of time explaining it wouldn’t train its AI on creators’ content).
Procreate vs. AI
Although it’s easy to see why creators would celebrate Cuda’s statement, there’s a tone in Procreate’s shriek against GenAI that sounds like whistling past the graveyard. This technology has been uncorked from its bottle, and there’s no going back. Certainly, there’s a lot that tech companies, creators, and the industry (indeed, every industry) need to resolve, but isn’t opting out all that essentially the same as giving up?
Not necessarily, at least not for a company like Procreate. This isn’t some brand new startup — it’s been around since 2011 and had earned its favored status among creatives long before generative AI was a thing. It’s not just paying lip service to the value of human creativity; it’s helped build it, and it has the receipts.
Putting aside Cuda’s salient point about generative AI’s effect on jobs and the economy (a very real, very worrisome, and much bigger topic), there’s little benefit for Procreate if they added generative features. In many, if not most cases, generative AI closes a skill gap: If you’re an untalented writer, a chatbot can give you text that passes for competent. But if you’re a good writer, anything AI gives you will be below a certain quality, and arguably just as much work to refine and edit as to write in the first place.
Similarly, the primary use case of AI features in creative platforms in Adobe Creative Suite, Canva, even video editors like Opus Clip is often a “democratization” of advanced features. Instead of spending hours figuring out how to do something, the software now allows users to just ask for what they want. The output of prompts, however, typically doesn’t match the results of those with true skill.
Moreover, actively working with the tools is how artisans refine that skill. It’s a little like getting a robot to lift weights for you: If the goal is doing as many reps as possible, then great. But if the goal is working the muscle, it makes no sense.
Creators Serving Creators
A better analogy for Procreate might be a restaurant where the patrons are all chefs themselves. It would be a poor choice for such an establishment to offer a “quick and easy” menu with items like chicken fingers and fries. But for a business with broader appeal — or looking to broaden its appeal — AI features can help advance that strategy.
That’s not to say producing mediocre content is AI’s only use case in creative apps. And certainly, there are non-public-facing uses of AI for a company like Procreate; I’m sure someone on its marketing or PR team has dabbled with ChatGPT or Claude (though if you haven’t, call me!).
But generative features don’t necessarily make sense for every product. There’s a reason “hand made” became a selling point after mass production became widespread. When your whole business revolves around celebrating the human element, robots — even competent ones — just get in the way.
The Media Copilot’s next AI Fundamentals course for journalists, marketers and PR professionals is coming Sept. 4. It’s a 100% live session where I walk creative professionals through how to use AI in their work.
AI Fundamentals for Media, Marketing and PR Pros (Wednesday, Sept. 4 @ 1 p.m. ET) is a loaded 3-hour class that takes you from structured prompting to advanced AI tools in a single afternoon. Did I mention the discount code? Apply the code AIHEAT at checkout for 50% off.
Why take a generic class that just gives you generalities on “mindset” when you can learn practical tools and techniques you can start using today?
The Chatbox
All the AI news that matters to media
Condé Nast Joins the Party: “Take the money” is becoming conventional wisdom among publishers who are being offered deals to use their content as food for the insatiable appetites of large language models. You can count the owner of Vogue, Wired, and Vanity Fair among them now that Condé Nast has signed with OpenAI. While SearchGPT is just a prototype right now, and its overall experience probably needs work, the value of the partnership to a large publisher like Condé is what it can learn — how often their articles are called in queries, what those queries are, and what revenue they might expect from other AI platforms. A lot is going to be figured out over the next year, and I’m starting to wonder if The New York Times is starting to feel left out.
Next Level AI Reporting: The Washington Post has a new AI tool for journalists: Haystacker applies AI to large datasets to better inform feature articles like this one, which analyzed more than 700 political ads for their visual content, extracting the meaning from on-screen text and visual indicators like the flag. Speaking to Axios, CTO Vineet Khosla said the Post would be building more tools in-house to better meet the needs of journalists. While most publishers abandoned the idea of being tech companies, AI presents a new approach to tool-building — one where prompt engineering goes almost as far as actual engineering. It might be time for the pendulum to swing back the other way.
Robot Reporters Outlawed at Ziff: Last month Ziff-Davis — the media company that owns PCMag, Mashable, and as of three weeks ago, CNET — agreed to unprecedented protections against AI replacing workers in its new agreement with the Ziff Davis Creators Guild, the editorial union of its publications. The new contract, ratified this week, prohibits union remembers being laid off or receiving a pay cut on account of AI. This is really a win for the right way to think about AI: If employers are seeing the tech properly — as assistive, and requiring human input in almost all use cases — then they’re not really giving up anything with such a provision, and it puts at least some of workers’ fears to rest. Other companies should emulate it.
Possible Futures for Journalism: Yes, we know journalism is in transition as the unprecedented power of AI slowly transforms the way we create and consume information. What we don’t know is what a fully transformed AI-mediated media ecosystem looks like. Technically, no one does, but the AI in Journalism Futures project from Open Society Foundations takes some of the best educated guesses we’ve seen. It’s a long read, though well worth it, with the report projecting five possible scenarios, including: one where journalists disappear as AI takes on the role of information “middlemen” between sources and consumers, a bifurcated future where some can pay for good information from AI and the rest get “noise,” and a highly regulated world that mitigates both the damage and the potential of AI. Glad we’re just speculating.
Midjourney for All: Speaking of AI progress, Midjourney — the AI-powered image creation service widely considered best-in-class — is now available to people with a low tolerance for Discord. The company’s web app is now available to everybody, freeing users of incessant audio notifications and slash commands. After kicking the tires on it, we’re pleased with how smooth they made the login and account-merger process, but generative imagery is as unreliable as ever. Case in point: Why is the head of George Washington’s horse smoking?
Leave a Reply