CLAUDE WAS USED IN THREE MISSILES PROJECT
According to Anthropic’s 154-page report, an armed structure in the north of Yemen used Claude in three separate missile projects. One of the accounts developed software while the others did research and technical control. The studies ranged from tactical rockets to ballistic systems targeting a range of over 2 thousand kilometers.
CLAUDE ANALYZED THE FAILED TEST
When the rocket test in the field failed, the test data was uploaded back to Claude. Artificial intelligence was asked to find the source of the malfunction and perform technical error analysis. Thus, it turned out that the system was used directly in a real weapons development cycle.
US NAVY WAS WATCHED
Anthropic announced that Iran-linked actors used Claude to collect information about US naval elements in the Middle East and to support military planning. Cases include espionage, mass surveillance and biological research.
CITY LIFE MAY BE LOCKED
According to experts, one of the biggest risks of advanced artificial intelligence systems is critical infrastructures. It is stated that if electricity, transportation, banking, communication, fuel and water systems are targeted simultaneously, life in big cities may come to a halt in a short time. Max Tegmark from MIT points out the rapid increase in artificial intelligence’s capacity to model biological systems. Experts are concerned that different biological threats may be produced at the same time and that superhuman systems may develop new attack methods against nuclear infrastructures.
Photos: AA, TRT
