Today’s frontier models create the problem.

That isn't actually relevant. They create problems, but when confronted with them and asked to fix them, my experience is that they can fix them very effectively. I haven't had an AI session get "stuck" on me, not being able to solve a problem in over a year.

The challenge as I see it is in shifting checks further left, but actual mitigation is just not really an issue any more.

What problem? They keep finding security bugs that humans couldn't for decades.

So ask them to fix the problems in a targeted way.