It's true that AI will be a tool used to harm people in a broad set of use cases, but the tools are too far removed from their respective applications in people's minds. People directly deal with health insurance companies, not the software vendors that sell the tools with sanitized descriptions.