OpenAI's Sora shutdown exposes deeper limits in AI's creative tools
OpenAI discontinued Sora, its video generation tool, on April 26, 2026. The shutdown wasn't sudden. The platform lost $1 million per day, failed to sustain user engagement after initial hype, and faced mounting legal constraints around copyright and intellectual property.
But the real story isn't about one failed product. Sora's collapse signals a pattern affecting the entire category of generative AI tools aimed at creative work.
The economics didn't work
Video generation demands far more computing power than text or image creation. OpenAI couldn't keep costs manageable or generate revenue that justified the expense. The company faced a choice: continue subsidizing a money-losing product or redirect resources to more profitable tools.
They chose the latter.
Users tried it once, then moved on
Sora launched in February 2024 to considerable fanfare. Early coverage called it a breakthrough. Yet traffic data tells a different story: users tried the tool, then stopped using it regularly.
The same pattern appears across the industry. Tools like Midjourney and Stability AI see initial spikes in adoption followed by declining sustained engagement. Creative professionals, in particular, haven't integrated these systems into their actual workflows.
AI's built-in bias against novelty
Here's the core problem: generative AI systems are optimized to replicate what they've already seen, not to create something genuinely new.
These models train on billions of images and videos selected for being polished, clear, and visually appealing. They learn patterns from existing work and get rewarded for matching those patterns closely. The result is outputs that look realistic but feel familiar-even interchangeable.
This creates what researchers call a "counter-creative bias." The systems suppress novelty in favor of familiarity. They produce work that's passable and palatable, which is exactly what creative professionals don't want.
Good creative work pushes boundaries. It breaks expectations. AI trained to imitate successful existing work can't do that by design.
Prompting is a poor substitute for artistic control
Using these tools requires learning a new skill: writing elaborate prompts with the right keywords and metaphors to coax the system into generating what you want.
OpenAI demonstrated this early on. When launching DALL-E 2 in 2022, the company showcased examples like "an espresso machine that makes coffee from human souls." The creativity came from the human-written prompt, not from what the AI generated.
Users spend time experimenting with prompt variations to reach acceptable results. They're not creating directly-they're manipulating language to control a system. For artists accustomed to working directly with their medium, this feels like an extra layer of friction, not a shortcut.
The internet is drowning in AI slop
Merriam-Webster and the American Dialect Society chose "slop" as their 2025 word of the year. The term describes the flood of low-effort, AI-generated images designed to drive engagement but devoid of creative substance.
Viral images of world leaders and wide-eyed children fill feeds. They're all variations on the same hyper-polished, generically pretty aesthetic. This is what happens when AI optimizes for familiarity at scale.
A different era of AI art is over
Between 2019 and 2022, artists experimented with AI as a genuine artistic medium. The Barbican Centre in London and the National Museum of China both showcased AI-integrated art. These artists trained models on their own work, not on billions of copyrighted images. The results were distinctive, experimental, and reflected each artist's vision.
Once major companies constrained AI image generation through language prompts and copyright safeguards, that exploratory period ended. The technology became less useful for serious artistic work and more suited to memes, spam, and automation.
For creative professionals, the lesson is clear: these tools aren't replacing your skills. They're solving a problem that doesn't match how you actually work.
Learn more about Generative Art and how to integrate AI effectively into creative practice through AI for Creatives.
Your membership also unlocks:
