Keeping a human in the loop is one part of AI agent resilience

None of the 19 enterprise respondents with AI agents in production allowed those agents to execute high-risk actions without human involvement, yet only four of the 25 technology and security leaders surveyed by theCUBE Research had tested an AI or agent incident scenario, principal analyst Krista Case said at the AI Cybersecurity and Resilience Summit. Eleven of the 25 practitioners acknowledged gaps in reconstructing agent actions across systems, and 12 had not addressed or planned to address detection of anomalous agent behavior, Case said, noting agents "can take hundreds or thousands of actions a minute." Organizations granting agents bounded autonomy over low-risk actions were more likely to invest in logging, monitoring and recovery infrastructure and to have tested incident scenarios, which Case said signals a more mature security program.

Keeping a human in the loop is one part of AI agent resilience Keeping a human in the loop is one way enterprises can govern AI agents’ actions. As those agents take on more work, security teams also need to consider how they would detect problems and recover from their effects. The relationship between AI trust and cyber resilience is becoming more complex as agents gain the ability to act across systems. Research conducted for the AI Cybersecurity and Resilience Summit https://siliconangle.com/2026/10/05/ai-trust-strengthens-cyber-resilience-digital-age-aitrustcyberresilience-aicybersecurityandresiliencesummit/ examined how organizations are governing those actions and preparing for incidents, according to Krista Case https://www.linkedin.com/in/krista-case/?utm source=chatgpt.com pictured, left , principal analyst and practice lead for cyber resilience and security at theCUBE Research. “We are actually leaning on human approval to secure and govern these AI actions,” Case said. “I think there’s a pretty big push and pull right now in the industry in terms of trying to figure out where we can start leaning on AI to move faster and at machine scale.” Case spoke with theCUBE’s Dave Vellante https://www.linkedin.com/in/dvellante/ right , co-founder and chief analyst at theCUBE Research, for the AI Cybersecurity and Resilience Summit https://www.thecube.net/events/thecube/ai-cybersecurity-and-resilience-summit , during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed how enterprises are governing, securing and recovering from the actions of autonomous AI agents. Disclosure below. Human in the loop and the recovery challenge The tension appears in theCUBE Research’s study https://thecuberesearch.com/ai-resilience-organizations-are-underprepared-and-largely-untested/ of 25 enterprise technology and security leaders. Of the 19 respondents with agents in production, none allowed them to execute high-risk actions without human involvement. Yet only four of the 25 had tested an AI or agent incident scenario. “It reminds me of the days when you wouldn’t test your disaster recovery. You might test failover, but you wouldn’t test failback,” Vellante said. “You’re seeing kind of a similar dynamic, but much more complicated today because you’ve got the state of the business that you have to recover as well.” Agent speed compounds the problem. Restoring an agent does not reverse the records, transactions and downstream workflows it already touched, which is why recovery may require reconstructing the sequence of actions it took, as Case wrote in theCUBE Research’s analysis of cyber recovery https://thecuberesearch.com/the-new-measure-of-cyber-recovery-can-the-business-get-back-to-work/ . Eleven of the 25 practitioners acknowledged gaps in their ability to reconstruct agent actions across systems, she noted. “The stakes are even higher, not only because it is the state of the business, but also because of the speed and scale that these AI agents are acting,” Case said. “They can take hundreds or thousands of actions a minute.” Bounded autonomy points the way forward One counterintuitive finding offers a path ahead. Organizations that let agents execute low-risk actions independently were more likely to invest in logging, monitoring and recovery infrastructure and to have tested agent incident scenarios, Vellante said. To Case, that bounded autonomy may indicate a more mature program. “If you are a security program that already has put some bounded autonomy in place for these AI agents … that indicates that you are a step further down the maturity path,” she said. Detection and forensics are the next areas to strengthen. Twelve of 25 respondents had not addressed, or were not planning to address, detecting anomalous agent behavior, and monitoring and reconstructing agent activity ranked as the least mature capabilities. “These are precisely the capabilities that are required if the human-in-the-loop controls break down,” Vellante said. Case expects progress to come quickly. “This space is very volatile. I think it’s going to be developing … very quickly,” she said. “Even just if we sit down at this point next year, I think we will be surprised at how far we’ve come.” Here’s the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of the AI Cybersecurity and Resilience Summit https://www.thecube.net/events/thecube/ai-cybersecurity-and-resilience-summit : Disclosure: Dell Technologies Inc. sponsored this segment of theCUBE. Neither Dell nor other sponsors have editorial control over content on theCUBE or SiliconANGLE. Photo: SiliconANGLE A message from John Furrier, co-founder of SiliconANGLE: Support our mission to keep content open and free by engaging with theCUBE community. Join theCUBE’s Alumni Trust Network , where technology leaders connect, share intelligence and create opportunities. - 15M+ viewers of theCUBE videos , powering conversations across AI, cloud, cybersecurity and more - 11.4k+ theCUBE alumni — Connect with more than 11,400 tech and business leaders shaping the future through a unique trusted-based network Are you an AWS customer? Support SiliconANGLE financially by buying your AWS services from our Marketplace portal page and links: https://siliconangle.com/aws-marketplace/ https://siliconangle.com/aws-marketplace/ About SiliconANGLE Media SiliconANGLE https://cts.businesswire.com/ct/CT?id=smartlink&url=https%3A%2F%2Fsiliconangle.com%2F&esheet=54119777&newsitemid=20240910506833&lan=en-US&anchor=SiliconANGLE&index=9&md5=646b1b564e2259100a2b8638aab0a552 , theCUBE Network https://cts.businesswire.com/ct/CT?id=smartlink&url=https%3A%2F%2Fwww.thecube.net%2F&esheet=54119777&newsitemid=20240910506833&lan=en-US&anchor=theCUBE+Network&index=10&md5=7de2a85f95ab4a4a495cede20b8cb1da , theCUBE Research https://cts.businesswire.com/ct/CT?id=smartlink&url=https%3A%2F%2Fthecuberesearch.com%2F&esheet=54119777&newsitemid=20240910506833&lan=en-US&anchor=theCUBE+Research&index=11&md5=7bb33676722925eb57d588ec343e4f6f , CUBE365 https://cts.businesswire.com/ct/CT?id=smartlink&url=https%3A%2F%2Fwww.cube365.net%2F&esheet=54119777&newsitemid=20240910506833&lan=en-US&anchor=CUBE365&index=12&md5=d310fb35919714e66ad8d42c9c0c1bc6 , theCUBE AI https://cts.businesswire.com/ct/CT?id=smartlink&url=https%3A%2F%2Fwww.thecubeai.com%2F&esheet=54119777&newsitemid=20240910506833&lan=en-US&anchor=theCUBE+AI&index=13&md5=b8b98472f8071b23ebb10ab9a8dd0683 and theCUBE SuperStudios — with flagship locations in Silicon Valley and the New York Stock Exchange — SiliconANGLE Media operates at the intersection of media, technology and AI. Founded by tech visionaries John Furrier and Dave Vellante, SiliconANGLE Media has built a dynamic ecosystem of industry-leading digital media brands that reach 15+ million elite tech professionals. Our new proprietary theCUBE AI Video Cloud is breaking ground in audience interaction, leveraging theCUBEai.com neural network to help technology companies make data-driven decisions and stay at the forefront of industry conversations.