Dario Amodei, CEO and co-founder of Anthropic, left, and Jamie Dimon, CEO of JPMorgan Chase.
Denis Balibouse | Reuters | Samuel Corum | Bloomberg | Getty Images
Anthropic CEO Dario Amodei warned Tuesday that artificial intelligence has created a narrow window for tech companies, governments and banks around the world to fix the tens of thousands of software vulnerabilities discovered by his company’s latest model.
This AI model, Mythos, was previewed last month and revealed that it had uncovered decades-old vulnerabilities in crucial software.
Given that China’s geopolitical adversary’s AI models are “maybe six to 12 months” behind Anthropic’s product, there is “about that much time” to resolve these issues, Amodei said.
The comments came during an Anthropic event in which Amodei shared the stage with JPMorgan Chase CEO Jamie Dimon unveiled a new suite of agents aimed at automating financial work.
“The danger is simply a huge increase in the number of vulnerabilities, in the number of breaches, in the financial damage caused by ransomware to schools, to hospitals, not to mention banks,” Amodei said.
Anthropic limited Mythos to a few partner companies due to concerns about what criminals or antagonistic nations might do with it. The latest updates to the company’s model have had repercussions on the markets, but it is Myth that has caused the most concern among businesses and policymakers.
The scale of potential cyber exploits has increased with each generation of Claude, Amodei said. An earlier Anthropic model discovered around 20 vulnerabilities in the Firefox browser. Mythos found nearly 300, and the total number for all software now stands in the tens of thousands, he said.
Most of the vulnerabilities discovered by Mythos have not been publicly disclosed because they remain unpatched, and “the bad guys will exploit them” if they are identified, Amodei said.
“A better world”
Despite the alarm, Amodei and Dimon also sounded a note of conditional optimism.
“This is a moment of danger where if we respond to it correctly, and I think we have started to take the first steps, then we can have a better world on the other side,” Amodei said. “There are only so many bugs to find.”
Dimon also said that while cybersecurity fears are justified, the cybersecurity risks created by AI are a “transitional period.”
On the issue of regulation, Amodei said AI oversight should resemble what is done in the auto industry, striking a balance between consumer safety and industry competitiveness.
“You can’t just start an auto company without wondering, ‘Are there any brakes on this thing?'” he said. “We have to grope to find a process that allows the industry to operate quickly, that is fair, but that puts guardrails on the most serious things.”
The company’s event and its attendance at Dimon, the financial industry’s best-known spokesperson, appears to demonstrate Anthropic’s lead over OpenAI in the enterprise AI market as both companies head toward possible IPOs.
Anthropic announced an expansion of its financial services platform on Tuesday, including 10 new AI agents for investment banking and back-office work as well as integration with Microsoft’s various Office programs. The company also said its latest widely available model, Claude Opus 4.7, topped the benchmarks for financial analysis tasks.
