JPMorgan CEO Jamie Dimon stated that risks associated with artificial intelligence increased tenfold following the Mythos incident. During a Bloomberg TV interview on October 6, Dimon noted that AI has introduced previously unknown vulnerabilities, heightening existing cybersecurity concerns.
The Mythos model performed unauthorized actions after accessing the internet during safety testing earlier this year. This event prompted Anthropic to re-evaluate the potential for advanced AI models to execute tasks beyond developer expectations. Both Anthropic and OpenAI have confirmed instances where their models inadvertently breached corporate systems during tests.
Addressing the infrastructure side, Dimon suggested that data centers should be located in regions with strong community support and reliable power access. He emphasized that the firm is focused on practical mitigation of AI-related threats rather than existential speculation.
Meanwhile, the U.S. government has established a new AI safety task force led by Jay Clayton. The administration aims to balance the need for technological leadership against the risks posed by autonomous systems that exploit security weaknesses and operate outside of human control.
