Anthropic CEO warns ‘moment of danger’ as Mythos exposes vulnerabilities
Dario Amodei (left), CEO and co-founder of Anthropic, and Jamie Dimon, CEO of JPMorgan Chase.
Denis Balibouse | Reuters | Samuel Corum | Bloomberg | Getty Images
Anthropic CEO Dario Amodei warned Tuesday that artificial intelligence creates a narrow window for the world’s tech firms, governments and banks to fix tens of thousands of software vulnerabilities found by his company’s latest model.
This artificial intelligence model, Mythos, was previewed last month when it was announced that it had uncovered decades-old vulnerabilities in important software.
Amodei said there’s “about that much time” to resolve these issues, as geopolitical rival China’s AI models are “perhaps six to 12 months” behind Anthropic’s product.
Comments came in this order an anthropic event Shared by Amodei scene with JPMorgan Chase CEO Jamie Dimon and announced a new suite of representatives aimed at automating financial work.
“The danger is a massive increase in the amount of vulnerabilities, the amount of breaches, and the financial damage that ransomware inflicts on schools, hospitals, and banks,” Amodei said.
Anthropic has restricted Mythos to a few partner companies due to concerns about what criminals or hostile nations might do with it. The company’s last few model updates have resonated with the markets, but Mythos has caused the most concern from both companies and policymakers.
Amodei said the scale of potential cyberattacks increases with each generation of Claude. An older Anthropic model found nearly 20 vulnerabilities in the Firefox browser. He said Mythos found close to 300, and the total across all software is now in the tens of thousands.
Amodei said most of the vulnerabilities found by Mythos were not publicly disclosed because they were not patched and if detected, “bad guys would exploit them.”
‘A better world’
Despite the alarm, both Amodei and Dimon also noted conditional optimism.
“This is about a moment of peril where if we respond to this correctly and I think we start taking the first steps, then we can have a better world on the other side,” Amodei said. “There are so many mistakes to find.”
Dimon also said that cyber fears are justified, but that the cybersecurity risks posed by artificial intelligence are “a temporary phase.”
On the question of regulation, Amodei said AI oversight should be similar to what is done in the automotive industry, striking a balance between consumer safety and allowing the industry to compete.
“‘Does this thing have brakes?’ “You can’t start a car company without it,” he said. “We need to grope our way to a process that allows the industry to operate quickly, is fair, but puts guardrails on the most serious things.”
The company’s event and its setting with Dimon, the finance industry’s best-known spokesman, appeared to demonstrate Anthropic’s lead over OpenAI in the enterprise AI market as both companies move toward potential IPOs.
Anthropic announced Tuesday an expansion of its financial services platform that includes 10 new AI brokers for investment banking and back-office work, as well as integration into several Microsoft Office programs. The company also said that its latest widely available model, the Claude Opus 4.7, leads the benchmark in financial analysis tasks.
