# AI Is Deskilling Site Reliability Engineers Through Automation Over-reliance on AI incident response threatens the critical human expertise required for complex system failures. - **Source:** [Hacker News](https://news.ycombinator.com) - **Published:** 2026-10-30 - **Topic:** [Software & Engineering](https://brifing.si/c/engineering) - **Original article:** - **On Brifing:** https://brifing.si/s/4880db21-d31d-40e6-b5e2-3eecd888aefc ## The Ironies of Automation Lisanne Bainbridge’s 1983 research paper remains highly relevant to modern software engineering. It highlights a paradox where increased automation makes human operators more critical during failures, yet simultaneously less prepared due to lack of practice. This creates a dangerous cycle where routine tasks are offloaded, leaving humans unable to manage rare, high-stakes crises. ## Routine Tasks vs Expertise AI currently handles alert noise reduction, log pattern detection, and basic auto-remediation for known failure modes. While these automations improve efficiency, they strip SREs of the routine cognitive tasks that build deep system intuition. By removing the need for manual triage, we inadvertently erode the muscle memory required for complex, novel system outages. ## The Deskilling Phenomenon Evidence from medicine and aviation shows that AI assistance leads to measurable skill degradation. Endoscopists using AI saw detection rates drop, while pilots relying on autopilot experienced reduced situational awareness. SREs face similar risks, as they stop reading raw logs or forming independent hypotheses, masking skill atrophy behind a facade of hyper-productivity. ## Never-Skilling for Juniors Beyond deskilling, the industry faces a never-skilling problem for junior engineers. In environments where AI resolves most incidents, new SREs lack the opportunities to build foundational intuition through hands-on practice. This broken training pipeline prevents the next generation from developing the gut-level understanding of system architecture necessary to navigate truly catastrophic, cascading failures. ## The Vicious Cycle SREs often recognize their skills are degrading and subsequently lean even harder on AI tools to compensate for their perceived incompetence. This creates a self-reinforcing cycle that is difficult to escape. Because AI makes tasks feel easier, users struggle to accurately judge their own performance until a novel incident exposes their lack of preparation. ## Deliberate Inefficiency Strategies To counteract skill decay, SRE teams should adopt deliberate inefficiency. Similar to FAA mandates requiring manual flying time for pilots, teams can designate specific incidents as human-practice opportunities. While this approach sacrifices speed for a faster resolution, it serves as a long-term investment in keeping human expertise sharp and ready for emergencies. ## Human-in-the-Loop Design Systems must be designed for human-in-the-loop engagement rather than human-on-the-side supervision. Active participation ensures engineers remain cognitively engaged with the problem-solving process. Approving AI recommendations is insufficient; engineers must actively work alongside AI to maintain the mental models and deep architectural knowledge required when automated systems fail to provide a solution. ## Observability and Data Context SigNoz leverages its position as an observability platform to build context-aware AI. By owning the underlying metrics, traces, and logs, the platform provides deeper insights than tools relying solely on external APIs. This approach aims to supercharge SREs by keeping them in control of production infrastructure rather than replacing their decision-making capabilities entirely. ## Future of SRE Roles The future of site reliability engineering depends on intentionally designing AI to support, not replace, human operators. Humans remain essential for the most critical production decisions. By prioritizing tools that keep engineers sharp and engaged, the industry can ensure that when novel, cascading failures occur, the human element remains a reliable, capable safeguard. ## Sources covering this story - https://newsletter.signoz.io/p/ai-isnt-replacing-sres-its-deskilling — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._