Resolution Criteria:
The Al's actions must be unexpected and not reasonably predictable by its operators. Incidents where operators were aware of the risks but ignored them do not count.
Deaths resulting from human decisions, where Al serves merely as a tool or assistant, are excluded (war, etc).
The question concerns the total number of deaths caused by Al. Multiple deadly incidents from different Al systems will be combined if it occurs within 365 days.
Background:
The potential for a rogue Al system to autonomously cause large-scale harm is a significant concern within the field of Al safety. This issue relates specifically to the Al alignment problem, which involves ensuring that advanced Al systems act in accordance with human values and intentions as they become more autonomous. Researchers at organizations such as the Machine Intelligence Research Institute (MIRI) have emphasized the dangers of failing to solve this alignment problem, warning that highly capable Al systems, if misaligned, could lead to potentially catastrophic outcomes.
1,000
3.00