I'm not too concerned short to medium term. I feel there are just too many edge cases and nuances that are going to be missed by AI systems.
For example, systems don't always work in the way they're documented to. How is an AI going to differentiate cases where there's a bug in a service vs a bug in its own code? How will an AI even learn that the bug exists in the first place? How will an AI differentiate between someone reporting a bug and a hacker attempting to break into a system?
The world is a complex place and without ACTUAL artificial intelligence we're going to need people to at least guide AI in these tricky situations.
My advice would be to get familiar with using AI and new AI tools and how they fit into our usual workflows.
Others may disagree, but I don't think software engineers (at least ones the good ones) are going anywhere.