Surely a lot of the load can be attributed to AI hallucinated submissions, but even then those need to somehow be reviewed on the off-chance that it's a real and important vulnerability - which would have usually been done by the human reporter in the past :)
