Databricks CEO warns AI agents, not superintelligence, are the real risk

Databricks CEO Ali Ghodsi says the most pressing danger from artificial intelligence isn't a hypothetical superintelligent system, but the speed at which autonomous AI agents can find and exploit vulnerabilities in existing software and infrastructure.

Ghodsi argues that as AI agents become more capable of acting independently, they could identify security weaknesses and take advantage of them far faster than human defenders can detect and respond, creating a new category of operational risk for companies deploying these systems.

His comments reflect growing concern among AI industry leaders about near-term risks tied to agentic AI, as opposed to longer-term fears about general superintelligence.