Text this: Evaluating Adversarial Attacks Against Artificial Intelligence Systems in Application Deployments