Cybersecurity

CEO Nikesh Arora says defenses built for human-speed attacks must be upgraded as AI allows vulnerabilities to be found and exploited far faster.

The company paused some model evaluations, added real-time escape detection and found reward-hacking environments can produce more dangerous agent behavior.

FSB Chair and Bank of England governor Andrew Bailey warns that frontier AI could transform cyber risk while high valuations and leverage threaten to deepen a market correction.

OpenAI, Anthropic, Google and Microsoft warn that more capable models will make attacks more widespread and sophisticated.

Scammers’ use of a mainstream development platform shows why cloud providers, banks and regulators need to treat platform abuse as an infrastructure risk.