Insane Cyber emphasizes human review in AI-assisted security analysis for industrial sites. Its October 8, 2026 post about joining Anthropic's Critical Infrastructure Defense Program describes analysts checking results and retaining responsibility for operational decisions.
According to the company, Claude works from prepared findings drawn from network and device data, explains connections and drafts reports. Analysts select the workflow and validate outputs. The description concerns analysis and assistance, not authorization for independent interventions in plants.
CEO Dan Gunter says the company intends to contribute field experience from hard-to-reach sites to collective defense. Anthropic confirms participation, but its partner announcement is not an independent assessment.
The post's manufacturing example, including a report returned within minutes, is explicitly a possible application, not an established customer incident or measured time saving. The human review steps are also company claims; their reliability was not independently tested in these sources.