Businesses ‘need to keep up with how agents behave,’ Wang says during the latest episode of CRN’s Security or Else! podcast. ‘How does that all work in this new world?’
As AI agents begin to work across a variety of browsers and IT systems, a massive challenge is arising for businesses whose identity security controls were designed around humans, according to 1Password CTO Nancy Wang.
In an interview for the latest episode of Security or Else!—a CRN video series hosted by senior editor Kyle Alspach—Wang said the stark differences between how agentic and human workers will operate is creating a major need for identity security to adapt across the industry and channel.
[Related: Netskope CEO Sanjay Beri: Rogue Agents Make An AI ‘Kill Switch’ Mandatory]
Unlike human employees, AI agents can appear—and disappear—within seconds, she noted. Agents can also operate in teams that have been assembled for a specific project and rapidly take “thousands upon thousands of actions.”
All of these dynamics creates a huge need for new thinking about how businesses will verify who is taking an action—and what that user ought to be sanctioned to do, Wang said.
“We simply can’t just apply the same deterministic controls for humans to agents, and it goes to how agents are created these days,” she said. “They’re created ephemerally.”
For solution providers and MSPs, the major shift also creates big opportunities to assist customers with understanding the fast-changing security market—as well as to make determinations about which protections best fit their environments, she said.
The reality is that the very nature of AI agents raises questions about what will happen to an agent’s identity when its container shuts down—and whether that identity should persist if the agent returns, according to Wang.
“Does the identity itself for that agent persist? Or does it go away when the agent or when the container is spun down? And does a new one get assigned when the container is spun up again?” she said.
“Those are all, I think, questions that we’ve never had to think about for humans. Because it’s not like a human goes away after a couple hours and comes back again, maybe a couple weeks later,” Wang said. “That just doesn’t happen in the human workforce.”
The challenge extends to granting access, as well, she noted. Zero-trust principles certainly remain important, but the systems enforcing zero trust will need to account for ability of agents to move at speeds far beyond those of a human worker, she said.
Ultimately, “your systems also need to keep up with how agents behave,” Wang said. “And if they’re assigned a human identity, but they work like machines, at machine speed, well, how does that all work in this new world?”
Check out the video here, and please like and subscribe if you enjoyed (available on YouTube, Spotify and Apple). Security or Else! features conversations with security experts and top executives about what’s hype and what’s real—and what channel partners truly need to focus on next in security and AI.
