Mass deployment of AI agents is a disaster, CertiK CEO warns

Ronghui Gu, co‑founder and CEO of CertiK, warned that the global rush to deploy autonomous AI agents is creating a catastrophic security debt. While companies market these tools as productivity boosters, unisolated and unvetted agents can be extremely risky when given broad access.

Agents are no longer just answering questions in a chat window: they call external tools, read local files, trigger workflows and interact with financial infrastructure. Gu cautioned that if execution environments are not isolated and tools are not scanned first, a compromised agent can gain broad internal access to a network.

Many popular open‑source AI applications assume that running locally or connecting via standard chat apps makes them safe, but Gu argued the opposite. Granting an agent permission to read system storage, view execution histories or manage credentials effectively turns it into the ultimate inside threat.

ronghui gu, certik, ai agents, autonomous agents, security debt, execution environments, isolation, credentials, open source, inside threat