AI Governance and Compliance on AWS: Config, CloudTrail, and Audit Manager

Task 5.1 is about securing the AI system — identity, encryption, keeping traffic private. This task is about governance and compliance: making sure the system is run according to policies and regulations, and being able to prove it. The two overlap, and several services here are ones a security engineer already knows, but the governance lens asks different questions: Is this resource configured the way our policy requires? Who did what, and when? Can we show an auditor evidence that we meet a standard?

A simple way to hold the domain together is the chain regulation to control to evidence. A regulation or internal policy says something must be true — for example, training data must stay in a particular country, or all activity must be logged. A control is how you make that true — a configuration rule, an encryption setting, a retention policy. Evidence is the proof you collected that the control is working — configuration history, API logs, a compliance report. Governance is running this chain continuously, not once.

For the exam, keep these services at a recognition level. You do not need to configure them; you need to know which one answers which governance question. The rest of this lesson takes them in pairs by the question they answer, gives you a comparison table, then covers the data-governance policies and processes that wrap around the tooling. Governance is as much about policy and process as it is about any single AWS service.

AWS Config answers is this resource configured the way it should be? It records the configuration of your AWS resources and how that configuration changes over time, and it can evaluate resources against rules that represent your desired state — for example, a rule that flags any storage bucket that is not encrypted. Config gives you a continuous picture of compliance: which resources meet your policies right now, which drifted, and what changed. For a generative AI deployment, Config can confirm that the buckets holding training data stay encrypted and that logging stays switched on. When a question asks how to track resource configuration and compliance over time, the answer is AWS Config.

Amazon Inspector answers a different question: are there known vulnerabilities in my workloads? It is an automated vulnerability management service that continually scans your workloads — such as compute instances and container images — for software vulnerabilities and unintended exposure, and reports prioritized findings you can act on. In a governance program, Inspector provides ongoing assurance that the infrastructure running your AI application is not carrying unpatched, exploitable software.

The pairing is easy to remember: Config watches configuration and compliance state, Inspector watches for vulnerabilities. Both run continuously so governance is an ongoing measurement rather than a one-time snapshot, and both feed the evidence you will later hand to an audit process.

Two services deal specifically with proving compliance against recognized standards, and the exam distinguishes them by whose compliance they document.

AWS Audit Manager helps you prove your compliance. It continuously collects evidence about how you use AWS and maps that evidence to the controls of a compliance framework — such as SOC 2, ISO 27001, PCI DSS, or a custom framework you define. Instead of manually screenshotting settings before an audit, Audit Manager gathers the evidence automatically and organizes it into an assessment, so preparing for an audit becomes far less painful. When a question mentions collecting audit evidence against a framework, the answer is AWS Audit Manager.

AWS Artifact gives you AWS's compliance documentation. It is a self-service portal to download AWS's compliance reports and certifications — such as SOC 1, SOC 2, ISO, and PCI reports — and to review agreements. These reports demonstrate how the AWS infrastructure underneath your workload is audited and certified, which is exactly the AWS side of the shared responsibility model that your own auditors will ask about. When a question asks where to download AWS's SOC 2 or ISO report, the answer is AWS Artifact.

The distinction is the exam's favorite trap: Audit Manager assembles evidence about your environment; Artifact hands you AWS's reports about AWS's environment.

AWS CloudTrail is your audit trail of API activity. It records the actions taken in your account — who made a call, which service and action, when, and from where — whether the caller was a person, an SDK, or another service. For governance this is the record of accountability: if someone changed a model's permissions, invoked a foundation model, or altered a data bucket, CloudTrail captured it. When an investigator or auditor asks who did this and when, CloudTrail has the answer. The exam trigger is direct: an audit trail of API calls means AWS CloudTrail.

It helps to separate CloudTrail from AWS Config, because they are often confused. CloudTrail records actions — the API calls that happened. Config records state — how a resource is configured and whether that configuration is compliant. A governance program uses both: Config tells you a bucket became non-compliant, and CloudTrail tells you which call changed it.

AWS Trusted Advisor inspects your account and provides best-practice recommendations across areas including security, cost optimization, performance, and fault tolerance. It flags issues such as overly permissive access or unused resources and suggests fixes. In a governance context, Trusted Advisor is a broad, ongoing checklist that surfaces where your account drifts from AWS best practices — useful hygiene alongside the more targeted compliance services.

These services often appear as answer options in the same question, so the discriminator is the exact governance job each one does. Read the scenario for the verb and the target.

A quick way to keep them straight: Config is about configuration state, CloudTrail is about actions taken, Inspector is about vulnerabilities, Audit Manager assembles your evidence, Artifact gives you AWS's reports, and Trusted Advisor is a broad best-practice checker. The exam rarely asks you to configure any of these. It asks you to pick the one whose purpose matches the sentence — so anchor on the purpose, not the details, and the confusable pairs (Config versus CloudTrail, Audit Manager versus Artifact) stop being confusable.

Beyond the tooling, governance is a set of decisions about how you handle data across its life. The exam lists several strategies, and each is a policy you set and then enforce.

Data lifecycles define how data moves through stages — created, used, archived, and eventually deleted — so nothing is kept forever by accident or thrown away too soon. Retention is the related rule for how long data is kept before it is deleted, often set by regulation; a policy might require training records to be retained for a set number of years, then removed. Data residency is about where data physically lives: many regulations require that certain data stay within a specific country or region, so you choose AWS Regions accordingly and prevent data from being processed elsewhere. For a generative AI system, residency can dictate where prompts, responses, and training data are stored and processed.

Logging records what happens to data and systems, and monitoring and observation mean actively watching those logs and metrics so you notice problems, misuse, or drift rather than discovering them after the fact. Put together, a data-governance strategy states the lifecycle and retention rules, pins data to compliant Regions for residency, and keeps everything logged and monitored. These are policy decisions first; the AWS services in this lesson are how you implement and prove them.

Governance also lives in how an organization operates. Policies are the written rules for building and running AI systems — what data may be used, who approves a model for production, what must be logged. A review cadence and defined review strategies mean these systems are checked on a regular schedule rather than set and forgotten, so a model is re-evaluated as data and risks change. Transparency standards require you to be clear about how AI is used and how decisions are made, and team training requirements ensure the people building and operating AI understand the policies and their responsibilities.

Governance frameworks give structure to all of this. The one the exam names is the AWS Generative AI Security Scoping Matrix — a framework for scoping the security of a generative AI workload based on how you use the model. It sorts generative AI use into scopes that range from simply using a public consumer application, through building on a pre-trained model, up to fine-tuning or training your own model. The point is that your security and governance responsibilities grow as you move up the scopes: someone using a public app carries far less responsibility than a team training a model on proprietary data. When you recognize the scope of your usage, you know which governance and security considerations apply. For the exam, recognize the Generative AI Security Scoping Matrix as a framework for scoping generative AI security — you do not need to memorize every scope.

Bring it together with a walkthrough. A healthcare company deploys a generative AI assistant that answers staff questions from internal clinical documents, and it must satisfy regulators.

First it writes the policies: which data the assistant may use, who signs off before it goes live, and a quarterly review cadence. For data governance, patient data must stay in-country, so the team pins storage and processing to AWS Regions in that country for residency, sets a multi-year retention rule with a defined data lifecycle, and keeps everything logged and monitored. To scope the work, it uses the Generative AI Security Scoping Matrix and identifies its usage as building on a pre-trained model with private data, which tells it how much governance responsibility it owns.

Then it wires up the AWS services. AWS Config continuously checks that data buckets stay encrypted and logging stays on. AWS CloudTrail records every API call as the audit trail of who accessed what. AWS Audit Manager collects evidence and maps it to the frameworks the company must meet, and AWS Artifact provides AWS's own SOC and ISO reports for the infrastructure layer. When the regulator arrives, the company can show the chain from regulation to control to evidence for each requirement — which is exactly what governance is meant to produce.