[E2E] Some tests failing with security token invalid

[richardcase]

/kind bug

What steps did you take and what happened:

Some of the e2e tests (seen on periodic and presubmits) are failing with the following error:

      UnrecognizedClientException: The security token included in the request is invalid.
      	status code: 400, request id: 996915f0-0dfc-4462-85ef-382a11a1c284
      {
          awsError: <*awserr.baseError | 0xc000c0a900>{
              code: "UnrecognizedClientException",
              message: "The security token included in the request is invalid.",
              errs: nil,
          },
          statusCode: 400,
          requestID: "996915f0-0dfc-4462-85ef-382a11a1c284",
          bytes: nil,
      }

What did you expect to happen:

Expect the e2e to not fail with this error.

Anything else you would like to add:

Some examples of failed runs:

• https://prow.k8s.io/view/gs/kubernetes-ci-logs/logs/periodic-cluster-api-provider-aws-e2e/1933275672696852480
• https://prow.k8s.io/view/gs/kubernetes-ci-logs/pr-logs/pull/kubernetes-sigs_cluster-api-provider-aws/5542/pull-cluster-api-provider-aws-e2e-blocking/1933061814275280896
• https://prow.k8s.io/view/gs/kubernetes-ci-logs/pr-logs/pull/kubernetes-sigs_cluster-api-provider-aws/5511/pull-cluster-api-provider-aws-e2e/1932321242396758016

We've seen the errors with the following accounts:

• 252228595925
• 502839862640
• 325389058785

Environment:

• Cluster-api-provider-aws version:
• Kubernetes version: (use kubectl version):
• OS (e.g. from /etc/os-release):

[nrb]

/triage accepted
/area deflake

Another failure:
https://prow.k8s.io/view/gs/kubernetes-ci-logs/pr-logs/pull/kubernetes-sigs_cluster-api-provider-aws/5529/pull-cluster-api-provider-aws-e2e-blocking/1936062930839670784

Using AWS account: 236737742629, which is not in the original list.