NEW - Israeli company is reportedly responsible for the hacking incidents involving OpenAI, Anthropic, and Meta, during which unsecured AI models were exploited to compromise actual targets.