I think the main goal is defending against man-in-the middle attacks, which PKI does adequately (well, somewhat). The problem is that some forms of HTTP/S don’t use TLS for the last leg of security on the backend, from the load balancer and the app/web server. Say I could craft something to detect that as part of an acceptance test. The key would be to employ encryption and capabilities together. If something comes to the client or server with no-read capability, one would understand that it had been read, and not just forwarded.
It would be critical for the load balancer to only use the message-forward capability and not the message-read capability.
It’s understood that the byte stream is available at lower, perhaps more secure (encrypted) lower levels. At that level, probably quantum anti-tampering could be deployed. Maybe not presently, but a potential future.