AI Is "Industrializing" Vulnerability Discovery: Anthropic Executive Warns Cybersecurity Risks May Worsen Further

An Anthropic PBC executive warned that large-scale cybersecurity disruptions, such as the situation this week in which governments and companies across multiple countries scrambled to patch two critical software vulnerabilities, could get "much worse" before they get better. This week, banks, airlines, and governments around the world took websites offline as a precaution to avoid being hacked. This followed cybersecurity researchers using AI to discover vulnerabilities in the open-source programming language PHP, which underpins a substantial portion of the internet. Jason D. Clinton, Anthropic's Deputy Chief Information Security Officer, posted on X on Thursday: "The past 4 days have already been bad, but it will get much worse before it gets better. Security leaders, don't waste any crisisgo to your board now and ask for help. Not for AI, but for the most basic security work."