Developed artificial intelligence systems can pose a threat to the existence of humanity, and the most powerful models can help create weapons of mass destruction. Developer Claude Anthropic warned about this in its IPO prospectus published on the website of the US Securities and Exchange Commission (SEC). The document states that particularly serious risks arise when artificial intelligence is used maliciously, becomes a weapon or when systems gain a high degree of autonomy. Anthropic acknowledges that its advanced models already have capabilities that could help attackers develop chemical and biological weapons. The company uses protection mechanisms, but warns that they can be bypassed.