Midjourney bans Stability AI from using its service: Here's why
Midjourney, a generative AI company, has imposed an indefinite ban on all employees of its rival firm, Stability AI, from using its service. The ban comes in response to allegations that Stability AI staff were involved in a data scraping attempt that led to a 24-hour system outage at Midjourney. The incident was first reported by the company on March 2, acknowledging an extended server outage, which prevented generated images from appearing in user galleries.
Midjourney links outage to Stability AI's "botnet-like activity"
Midjourney identified the cause of the outage as "botnet-like activity from paid accounts," which it linked to Stability AI employees. The company made this claim during a business update call on March 6. Midjourney user Nick St. Pierre, who attended the call, reported on X that the company accused someone at Stability AI of trying to extract all prompt and image pairs during a late-night operation. X user @BuckedUnicorn stated that there were two accounts tied to one individual.
Stability AI CEO investigates allegations, denies intentional disruption
On March 7, Stability AI's CEO, Emad Mostaque, responded to the allegations by stating that he was looking into the matter. He denied that his company had ordered any such actions and suggested that if a Stability employee was responsible for the outage, it was unintentional. Mostaque also dismissed claims of a DDoS attack and expressed confusion over how two accounts could cause such disruption.
Midjourney introduces new policy against 'aggressive automation'
In response to the recent outage, Midjourney has announced a new policy that will ban workers of any company found to be engaging in "aggressive automation" or causing service disruptions. This policy was revealed during the same business update call on March 6, where the allegations against Stability AI were discussed. The ban on Stability AI employees is part of this new policy implementation.
Midjourney and Stability AI face criticism over data usage
Both Midjourney and Stability AI have faced criticism from creatives for their use of large amounts of online data, allegedly scraped without consent, to train their models. The irony of the recent incident has not been lost on observers. Both companies have been targeted with several copyright lawsuits, with Midjourney being accused in December of creating an artist database for training purposes.