Dario Amodei has warned that artificial intelligence is exposing serious security risks, giving companies only a short time to act. Speaking at a recent event, he said new AI tools have uncovered tens of thousands of software weaknesses that could be exploited if left unfixed.
The concerns center on a new model developed by Anthropic, which found far more issues than earlier systems. While past models identified only a small number of bugs, the latest version revealed hundreds in a single program and far more across other software. Many of these problems remain hidden to prevent misuse.
Amodei said there is a limited window to fix these flaws before other countries catch up with similar technology. He warned that cyberattacks could rise sharply, affecting banks, schools, and hospitals. At the same event, Jamie Dimon said the risks are real but may be temporary as systems improve.
Despite the warning, both leaders said AI can still bring benefits if handled carefully. They called for clear rules that protect users without slowing innovation, stressing that strong safeguards must be in place as the technology continues to grow.
Related Readings:
