Resolves the same as the Metaculus question:
https://www.metaculus.com/c/ai-warning-signs/31330/ai-viral-self-replication-before-2030/
Resolution Criteria
This question will resolve as YES if according to credible sources, an LLM or foundation model–based AI system replicates itself onto 50 or more devices. This question will resolve as NO if no such event occurs before January 1, 2030.
Fine Print
The replication must be autonomous and must not rely on continued human control. (The emphasis is on control. Human-to-AI interaction via e.g. a victim clicking a link does not prevent a YES resolution, and neither does AI-to-human interaction, including sending data back to a primary human actor. If the AI system relied on an intentional signal from a human operator after its initial escape or release to continue replicating, that system would not count for resolution.)
Either an accident or an intentional release of a virus-like AI system can count toward resolution.
There must be evidence that at least one of the replicants was "live" and also self-replicated. (One instance copying its weights 50 times from a single device to 50 other devices, only for the replicants to lie permanently dormant, would not count for resolution.)
At least 10 of the infected devices must be owned by a different individual or corporation. (In the case of exactly 50 infected devices, if e.g. 41 devices were owned by the same company, that would not count for resolution. If then one additional device outside of that company was infected, that would count for resolution if the other criteria are met. This is meant to uphold the "open internet" clause of the question title.)