This is why such questions always relate back to the debate of automation vs manual flying ability.
And i guess both options will be separate until we can safely say technology can take over what a human brain would have to do, in emergency responses, checklists, quick analysis and reaction times and much more that we cannot yet automate. But at that time, the world of pilots will not be the good world where we can still fly planes.
Back to reality, we cannot assume that TCAS should not be merged with autopilot due to a chance it could fail. Fail safes should always be built to prevent disasters (while not talking about Boeing and its MAX).
I guess its simpler to maintain both separate and let the pilot handle the advisory, but could change in the future, depending on plane configurations (more automation or maintain certain manual controls to keep attention to safety)