Skip to content
IFAI

Industry

Midjourney bans rival Stability AI over image scraping

Published Reach 42

Midjourney just slapped an indefinite ban on all Stability AI employees after bot-like scraping brought its service down for 24 hours. The rival was allegedly trying to harvest prompt-image pairs to train its own models. Now, Midjourney is enforcing a nuclear policy: crash our servers with aggressive automation, and your entire company gets banned.

Midjourney bans rival Stability AI over image scraping
Image: arstechnica.com

There is a special kind of irony in watching an AI image company lose its mind because someone scraped its images.

Midjourney dropped an indefinite ban on every single employee working at rival Stability AI. The announcement came through Midjourney’s official Discord channel and was quickly shared on X by longtime community advocate Nick St. Pierre.

Midnight scrapers and a 24-hour crash

The drama started around midnight on March 2, when Midjourney’s infrastructure got hammered by massive, botnet-like traffic. The onslaught took the commercial image generator down for a full 24 hours.

When Midjourney investigated what happened, it tracked several paid accounts responsible for the surge back to an employee on Stability AI’s data team. The goal was straightforward: harvest large volumes of prompt and image pairs, the exact fuel needed to train or fine-tune rival models like the upcoming Stable Diffusion 3.

Midjourney did not just ban the offending accounts; it banned the entire company indefinitely. It also rolled out a brand-new policy stating that any company using aggressive automation that knocks Midjourney offline will see all of its employees banned on the spot.

The ultimate copyright irony

The internet wasted no time pointing out the elephant in the room. Siobhan Ball of The Mary Sue remarked how rich it was for a company built entirely on scraped web data without artist consent to complain about being scraped, saying cue the world's smallest violin.

Human artists felt the irony even deeper. Photographer and artist Jingna Zhang pointed out in a viral post how dehumanizing it was to see her name used more than 20,000 times as a prompt in Midjourney, reducing her life's work into fodder for a commercial image slot machine.

Midjourney happily charges monthly subscriptions for images built on the backs of millions of uncredited human creators, yet the moment a competitor uses an automated script to grab those very outputs, Midjourney treats it like high-level sabotage.

Why it matters

For paying users, this outage is a harsh reminder of platform risk. When you build daily creative workflows around a service that runs through Discord and centralized servers, an aggressive scraping bot can wipe out your production pipeline for an entire day.

For the AI sector, this incident exposes how desperate top labs are getting for synthetic training data. High-quality prompt-image pairs are pure gold for fine-tuning newer models. The fact that Stability AI staff allegedly resorted to scraping a direct rival suggests the competitive gap is widening and organic data reserves are drying up.

We think Midjourney’s blanket ban makes sense from a server defense standpoint, but its moral stance is laughably hollow. You cannot build an empire on uncompensated scraping and then pretend scraping is a criminal sin when it cuts the other way. As synthetic data becomes the main battleground, these double standards will only make frontier labs look more hypocritical to the public and the courts.

Sources (1)

  1. [1]
    arstechnica.com(opens in a new tab)Originalarstechnica.com

Best of AI, daily