JPMorgan Chase CEO Jamie Dimon warned on Tuesday that while artificial intelligence holds promise for strengthening future cyber defenses, it is currently exacerbating security vulnerabilities. Speaking during the bank's earnings call, Dimon noted that AI tools are making it easier for malicious actors to exploit software, thereby creating new risks for financial institutions. He confirmed that JPMorgan is actively testing Anthropic's newly announced Mythos model as part of its strategy to harness AI benefits while mitigating threats from adversaries using the same technology. Dimon highlighted that the recent capabilities of AI have shifted the landscape of cybersecurity. He stated that AI has made threats more severe and complex, noting that while better defensive mechanisms may emerge later, the immediate impact is increased exposure. His comments followed reports that Anthropic's Mythos model had already identified thousands of vulnerabilities in corporate software, a discovery that underscores the urgent need for immediate fixes. Earlier this week, Treasury Secretary Scott Bessent convened a meeting with bank CEOs to discuss the specific risks posed by the Mythos model, reflecting growing governmental concern. Despite JPMorgan's significant investments in cybersecurity, including dedicated expert teams and constant coordination with government agencies, Dimon cautioned that no single institution is entirely immune. He emphasized that the financial system is highly interconnected, meaning that weaknesses in exchanges or other linked entities can create cascading risks. Jeremy Barnum, JPMorgan's Chief Financial Officer, reinforced this point by explaining that AI tools function as a double-edged sword. While they help identify flaws, they can also be deployed in attack mode by bad actors. Recent advancements from Anthropic and other developers have simply accelerated an existing trend rather than created a new one. In addition to discussing advanced AI threats, Dimon stressed that foundational cybersecurity practices remain critical. He argued that basic hygiene, such as securing data, protecting networks and hardware, and regularly updating passwords, still dramatically reduces risk. He suggested that sophisticated AI tools cannot replace the necessity of these fundamental measures. Meanwhile, other major financial players are also assessing the technology. Goldman Sachs CEO David Solomon confirmed on Monday that his bank is testing the Mythos model, though he declined to provide further details. The remarks from both leaders illustrate the cautious approach the banking sector is taking toward integrating AI, balancing its potential productivity gains against the immediate and tangible dangers it poses to system security.