In a recent development, Midjourney, a generative AI company, has taken action against Stability AI by banning its employees from using its services. The move comes in response to an outage that occurred earlier this month, which Midjourney claims was caused by Stability AI staff attempting to scrape data from their platform.

The server outage, acknowledged by Midjourney on March 2nd in an update on its Discord server, resulted in a disruption of generated images appearing in user galleries. Midjourney subsequently conducted a business update call on March 6th, during which they attributed the outage to "botnet-like activity from paid accounts," specifically linking these accounts to Stability AI employees.

According to a user named Nick St. Pierre, who listened to the call, Midjourney revealed that the service was brought down when someone from Stability AI allegedly attempted to gather all the prompt and image pairs in the middle of a Saturday night. The company reportedly connected multiple paid accounts to an individual on the Stability AI data team.

As a response to the incident, Midjourney has taken the step of banning all Stability AI employees from using their service indefinitely. Additionally, they are implementing a new policy that will extend similar bans to employees of any company engaging in "aggressive automation" or causing outages to their service.

St. Pierre notified Stability AI CEO Emad Mostaque about the accusations, who responded, stating that he was investigating the situation and denied ordering the actions described. Mostaque emphasized that Stability AI had been using synthetic and other data, insisting that if a Stability employee was involved, it was unintentional and not a deliberate DDoS attack.

Midjourney founder David Holz chimed in, offering assistance in the investigation and providing Mostaque with relevant information. While the situation is still unfolding, no additional updates have been provided since their conversation on March 6th. As of now, both Midjourney and Stability AI have yet to respond to requests for comment from The Verge.

The circumstances surrounding the alleged scraping incident raise questions about the extent of the server outage caused by just two accounts. The irony of the situation has not gone unnoticed by online creatives, who have been critical of both companies, as well as generative AI systems in general, for training their models on scraped online data without consent. Both Stability AI and Midjourney have faced multiple copyright lawsuits, with Midjourney even being accused of building an artist database for training purposes in December.

The aftermath of this latest incident will undoubtedly be closely monitored by industry observers and intellectual property rights advocates alike, with implications for the future of generative AI and data usage ethics.