Anthropic CEO Warns US Firms of Narrow Window to Fix Vulnerabilities Found by AI Model Mythos
Anthropic CEO Warns US Firms on AI-Discovered Flaws

Dario Amodei, CEO of AI company Anthropic, has issued a stark warning to US companies, banks, and government agencies, stating they have only a narrow window of time to address tens of thousands of software vulnerabilities uncovered by the company's latest AI model, Mythos.

Vulnerabilities Uncovered by Mythos

According to a report by CNBC, Amodei made these remarks during an Anthropic event where he shared the stage with JPMorgan CEO Jamie Dimon. Previewed last month, the Mythos model sent shockwaves across the financial and technology sectors by revealing decades-old flaws in critical software. Amodei explained that while earlier Anthropic models found dozens of vulnerabilities—such as 20 in Firefox—Mythos uncovered nearly 300 in the same browser. The total count across all software now runs into the tens of thousands.

Most of these vulnerabilities remain undisclosed to prevent exploitation, but Amodei cautioned: "The danger is just some enormous increase in the amount of vulnerabilities, in the amount of breaches, in the financial damage that's done from ransomware on schools, hospitals, not to mention banks."

Wide Pickt banner — collaborative shopping lists app for Telegram, phone mockup with grocery list

China's AI Timeline

Amodei also emphasized urgency, noting that AI models from China are only six to twelve months behind Mythos. This, he said, gives US institutions "roughly that amount of time" to patch vulnerabilities before adversarial actors could weaponize similar capabilities.

New AI Agents and Integration

Anthropic also unveiled 10 new AI agents designed to automate financial work, from investment banking to back-office operations, and announced integration across Microsoft Office programs. The company highlighted that its latest widely available model, Claude Opus 4.7, leads benchmarks for financial analysis tasks.

Regulation and Oversight

On regulation, Amodei compared AI oversight to the automotive industry: "You can't just start a car company without 'Are there brakes on this thing?' We need to grope our way to some process that lets the industry operate expeditiously, is fair, but puts guardrails on the most serious things."

Pickt after-article banner — collaborative shopping lists app with family illustration