Artificial intelligence companies Stability AI and Midjourney Inc. are fierce competitors in the AI image generation space. The rivalry has allegedly boiled over into the dangerous and felonious realm of corporate espionage.
Last weekend, Midjourney experienced a 24-hour outage that the company believes resulted from “botnet-like activity from paid accounts.”
These sorts of attacks are common culprits for various online outages, like websites going down or server-based services being interrupted. A conviction of a DDoS attack carries with the risk of a 10-year prison sentence. Even conspiring to commit this attack could land a person behind bars for up to five years.
However, even with potentially devastating punishments, DDoS-related outages are not uncommon. Of course, outages due to technical errors and bad luck occur too. Not every outage is the result of malice. So, what makes this particular Midjourney outage noteworthy?
According to AI industry expert and Midjourney evangelist Nick St. Pierre, Midjourney’s “Office Hour,” which are company news recaps for Midjourney employees and users, allege that Stability AI employees are behind the Midjourney outage. Midjourney also believes the incursion was an effort to scrape prompt and image pair data from Midjourney.
In MJ office hours they just said someone at Stability AI was trying to grab all the prompt and image pairs in the middle of a night on Saturday and brought down their service.
MJ is banning all of the stabilityAI employees from Midjourney immediately
This is breaking now
— Nick St. Pierre (@nickfloats) March 6, 2024
Per Alie Jules on X, formerly Twitter, Midjourney’s outage began “in the middle of the night” and was linked to an email account from Stability AI.
Stability AI’s founder and CEO, Mohammad Emad Mostaque, known as Emad, has denied the allegations on X.
“If anyone did do this on team (have asked, will dig, also happy if MJ reaches out direct) its not great but obviously not a DDoS attack but unintentional,” Emad added in a later reply on X. “Certainly not instructed by us/Stability AI tho, really happy with our dataset & the augmentations we have on that.”
David Holz, Midjourney’s CEO, jumped into the conversation as well, adding that he had sent Emad “some information” that would help Stability AI with its international investigation into the claims.
The companies behind many of the major AI models are no strangers to the concept of scraping someone else’s data, of course. That’s how AI models are built and trained. However, any sort of malicious cyberattacks are a totally different situation, and punishable by serious financial penalties and even prison time.
This situation is still developing as Midjourney and Stability AI do further digging into what precisely happened to Midjourney last week. The only proof that any Stability AI employee was involved comes from Midjourney itself, and of course, Stability AI denies any knowledge of such an attack.
In the fallout, Midjourney has allegedly banned all Stability AI employees from the service.