Revealing Mythos is tantamount to handing over ballistic missiles? JPMorgan Chase CEO makes shocking statement warning of risks
On Wednesday, JPMorgan CEO Jamie Dimon warned at the Defense and Innovation Summit in Pennsylvania that the risks posed by Anthropic's Mythos artificial intelligence model were a "real issue." He even believes that opening the Mythos model to the public is tantamount to handing over ballistic missiles to the public. Dimon expressed the "real problem" in May this year. The number of software vulnerabilities discovered by Mythos had reached tens of thousands at that time, and most of them had not yet been repaired. It was expected to take 6 to 12 months to be fully repaired.
On Wednesday, JPMorgan CEO Jamie Dimon warned at the Defense and Innovation Summit in Pennsylvania that the risks posed by Anthropic's Mythos artificial intelligence model were a "real issue." He even believes that opening the Mythos model to the public is tantamount to handing over ballistic missiles to the public. Dimon expressed the "real problem" in May this year. The number of software vulnerabilities discovered by Mythos had reached tens of thousands at that time, and most of them had not been fixed. It was expected to take 6 to 12 months to fully fix. Anthropic publicly disclosed the Mythos large model in April this year, and officially released Mythos 5 and Fable 5 on June 9 this year. Among them, Mythos 5 is a top-notch large model that can break through network security barriers; Fable 5 is a version with a network security protection lock and is currently open to the public. In June of this year, the U.S. government asked Anthropic to restrict access to Mythos 5 and Fable 5 to non-U.S. citizens due to national security concerns. After 18 days of security coordination and review, the U.S. Department of Commerce lifted restrictions on the model. The U.S. government is still paying close attention to this large model. Jamie Dimon has said that while artificial intelligence tools can help companies defend against cyberattacks, it first makes them more vulnerable. It is worth noting that one of the reasons why the Mythos model has attracted much attention from banks is that it is currently one of the most advanced artificial intelligence for identifying network security vulnerabilities, and can quickly help enterprises discover and repair security vulnerabilities. At the same time, this also means that after the hacker obtains the model, the cost of breaking into the bank gateway to obtain payment, credit and customer data is reduced. Anthropic has admitted this in media interviews, which is why the Mythos model is only open to a few institutions. However, industry insiders said that before Mythos entered the public eye, the industry had too many network security vulnerabilities that were too late to be repaired, and hackers themselves had such capabilities. At the same time, some professionals believe that artificial intelligence has changed the industry landscape, but many financial institutions still rely on old technologies and supplier ecosystems. Among them, hacker attack surfaces are mainly concentrated on core processing systems, digital banking service providers and loan issuance platforms. Therefore, in terms of response measures, in order to keep up with the pace of artificial intelligence, companies need to hire more cybersecurity experts and tighten control over third-party software security. As one of the few institutions approved to use Mythos since April this year, JPMorgan Chase has deployed hundreds of employees full-time to work on system security reinforcement. As early as last year, it announced that it would invest US$1.5 trillion in related industries to strengthen the security and resilience of the US economy over the next ten years. (