Will AI replace cybersecurity jobs? A cloud skills perspective

If you are considering a technical career, the question “will AI replace cybersecurity jobs?” often sits behind a more practical one: which skills will let me do useful work as the tools change?

For cloud and DevOps learners, security is already part of that work. A deployment needs an identity. An application needs appropriate access to data. A failed request needs enough evidence to investigate. Understanding those connections gives you a strong place to start.

How is AI changing security work?

An assistant can summarise an alert, explain a policy or suggest questions to investigate. It can help you work through unfamiliar information faster. You still need to establish whether the explanation matches the system you are looking at.

A useful distinction is between a task and a role. The ILO's April 2026 brief on AI exposure explains that measures of what AI could automate do not, on their own, predict employment outcomes.

For your learning plan, concentrate on work you can practise and demonstrate: understand an application's access needs, test a control, read the evidence and make a measured change. These skills fit directly into cloud engineering.

Where cloud, DevOps and security meet

Cloud engineering covers the environment an application uses. DevOps connects its development, delivery and operation. Security helps ensure that access and changes are appropriate throughout that process.

You encounter all three when deploying a service on AWS. You configure where it runs, automate how it is released and decide what it can access. You do not need to postpone security until after the application is finished; small, sensible controls belong in each stage.

The value comes from connecting a finding to an application you understand. A list of alerts is only a starting point.

A practical lab: give an application the access it needs

Use an isolated AWS lab and synthetic data. The example is a small service that reads one test object from storage and returns a short result. The learning objective is to demonstrate both allowed and denied access.

Define the intended behaviour

Write two sentences before configuring anything: “The service may read this test object. It may not read other test objects or write to storage.” That gives you a simple specification to test.

Decide which identity the application will use. When it runs on a suitable AWS service, use the service's role mechanism rather than copying long-lived credentials into application code.

Configure a narrow permission

Create the policy around the required action and resource. Keep the test resource private and avoid broad permissions just to make an error disappear.

AWS's IAM security best practices describes temporary credentials and least-privilege access. Apply the principle to your small lab: the service gets the permission it needs for the documented task.

Test the boundaries

Make the allowed read request and record the result. Then test an unrelated read and a write, and verify that they are denied. Keep the test outputs with your project notes without exposing credentials or private data.

An access error is useful evidence in the denied tests. It shows that the boundary behaves as intended. If a denied test succeeds, inspect the actual identity and all permissions affecting the request before changing the policy.

Add evidence you can investigate

Record the application's request identifier, action and result. Avoid logging credentials or sensitive payloads. If your investigation depends on AWS audit events, confirm that the relevant event type is enabled: resource-level data events require explicit configuration and can incur charges.

Now deliberately remove the permitted action in the lab. Observe the failure, identify the cause and restore the narrowly scoped permission. Write down how you distinguished an access failure from a network or application problem.

Keep the change reproducible

Store the infrastructure configuration and tests in Git, with secrets excluded. Include setup and cleanup steps so another learner can repeat the exercise.

The finished project demonstrates AWS, access control, testing and troubleshooting together. It is a useful cloud engineering story to discuss in an interview.

Build security checks into the release process

Once the application works, connect the controls to its delivery process. Review changes to permissions, keep application tests in the pipeline and restrict access to deployment credentials.

Use dependency checks to identify updates that need attention, then test the resulting change. Keep a record of the version you deployed so an investigation can connect application behaviour to a release.

Our GitHub Actions security checklist covers workflow concerns in more detail. The guide to AI and DevOps engineering shows how permissions, release checks and recovery fit into one project.

What should you learn first?

Start with Linux, networking, Git and Python, then learn how to deploy a small application on AWS. This gives you the context to understand identities, credentials, network boundaries and logs.

Add containers, CI/CD and infrastructure as code as you develop the project. Practise investigating a failed request and explaining the decision behind each control. If you already work in IT support or administration, troubleshooting and access-management experience can be useful foundations to build on.

From there, you can deepen your cloud and DevOps work or explore a security specialism. Our AI-era career guide compares related paths and the practical work behind them.

Learn secure cloud delivery with CloudPros

CloudPros teaches the skills together: Linux and Python, AWS, containers, deployment automation, infrastructure and monitoring. Security has a practical place within that work, from giving a service appropriate access to reviewing the changes that reach an environment.

Live classes, projects and instructor feedback help you develop an understanding you can explain and apply. The 20-week programme includes 16 core teaching weeks, three dedicated recap weeks and a bonus MLOps week, alongside CV, LinkedIn and mock interview support.

Explore the CloudPros curriculum for the full programme and read student reviews to hear how graduates describe their experience. To see how the same foundation supports model-powered applications, read the future of cloud computing.

Questions about AI, security and cloud skills

Will AI replace cybersecurity jobs?+

AI can automate or assist parts of security work, such as summarising alerts and checking configuration. Roles also involve understanding business systems, verifying evidence, choosing controls and managing incidents. Cloud access, deployment and investigation skills provide a practical foundation for this work.

Is cloud security part of cloud engineering and DevOps?+

Yes. Cloud and DevOps engineers work with permissions, secrets, network access, deployment checks and logs. Specialist security teams may own additional responsibilities, while engineers apply security controls as they build and operate applications.

What is a good beginner cloud security project?+

Deploy a small application in an isolated AWS lab with permission to read one test resource. Verify that the allowed read succeeds and an unrelated read or write is denied. Keep credentials out of Git, inspect the relevant logs and document how the controls work.

What should I learn before DevSecOps?+

Learn Linux, networking, Git, scripting and a cloud platform, then practise containers and CI/CD. Understanding the application and its delivery process makes it easier to apply useful checks, limit access and investigate errors.

How does CloudPros teach security skills?+

CloudPros teaches security in the context of cloud and DevOps work, including AWS access, application delivery, infrastructure and monitoring. The programme focuses on building and running cloud systems, with live teaching, practical projects and instructor feedback.