Skip to main content
Question

How is AI Changing Cloud Security Operations in 2026?

  • July 29, 2026
  • 2 replies
  • 38 views
aiworkflowindia
Forum|alt.badge.img

Hello Everyone,

With the rapid adoption of AI across organizations, I'm interested in understanding how security teams are adapting.

Some questions I'd love to hear your thoughts on:

  • Are your security teams using AI tools in daily operations?
  • Which AI tools have actually improved productivity?
  • What are the biggest security concerns when employees use tools like ChatGPT, Gemini, or Copilot?
  • Has your organization created an AI usage policy?
  • Do you believe AI will reduce manual security work or simply change the nature of it?

In my experience working with professionals and organizations on AI adoption, I've noticed that many teams are excited about AI but are also concerned about data privacy, compliance, and governance.

I'd love to learn from this community:

  1. What AI tools are part of your security workflow today?
  2. What challenges have you encountered?
  3. What best practices would you recommend for organizations just starting their AI journey?

Looking forward to hearing your experiences and learning from the community.

Thank you!

    2 replies

    a_aleinikov
    Forum|alt.badge.img+8
    • a_aleinikovBronze 2
    • Bronze 2
    • July 29, 2026

    AI is already useful in security operations for alert summarization, investigation support, log analysis, and playbook generation. The main risks are sensitive data exposure, weak access controls, and overreliance on unverified output. Organizations should define an AI usage policy, restrict approved tools, prevent confidential data from being submitted, and keep analysts responsible for final decisions. AI will reduce repetitive work, but it will also make validation, governance, and threat modeling more important.

    Forum signature: Aleksei Aleinikov | Senior Platform Security Engineer | Google Cloud | Security | IAM | Kubernetes | Terraform
      whathehack81
      Forum|alt.badge.img+9

      The biggest risk I see is not the AI tool itself, but how casually sensitive operational data gets pasted into it. Teams need clear boundaries around what data can be submitted, which models are approved, whether prompts and outputs are retained, and how generated results are validated.

      I use AI mainly for accelerating analysis, drafting detection logic, summarizing technical material, and organizing investigation steps. It helps with speed, but it does not replace analyst judgment. The output still needs to be tested against the actual environment, especially when it affects detections, response actions, or access decisions.

      For teams just starting, I would begin with low-risk internal use cases, define an acceptable-use policy, log usage where possible, and require human review for anything operational. AI will reduce repetitive work, but it will also make validation, governance, and threat modeling more important.

       

      I often have engagements, that request that I divulge how I will implement AI in to my workflow.  I tell them the most important rule, is to realize. Just because something "looks interesting"  does not mean it is a valid security concern. Only human validation, can determine; impact, and blast radius. First it was just the "HaCkEr". The only advisory. Now it's AI+¥ Hacker+Intelligence-by_automation. We Now have the attacker, that can just sit back and wait for his AI, to validate his attack path. 🤔 WHAT THE ????????  Makes me want to run to the box and curl -sS -v -i "https://can someone tell me what happened..wtf" | grep  -i "clarification".     🤣 

      R.Q whathehack81
      Terms & ConditionsAccessibility statement
      Privacy Policy Terms of Service Gen AI Policy Community Guidelines Cookie Settings

      © 2025 Google. All rights reserved.