Yes, that’s why I wrote “in theory”. In reality, it isn’t that clean, even with modification.

Integration may not just be occurring at AAVS1, but all over the place. Studies have detected things like complex vector rearrangements and truncated vector genomes across multiple animals models inserted around transcriptional units. There seems to be no preference for gene coding regions and no clustering of integration sites.

In fact, The FDA had a CTGTAC meeting in 2021 to discuss these issues: https://www.fda.gov/advisory-committees/cellular-tissue-and-gene-therapies-advisory-committee/2021-meeting-materials-cellular-tissue-and-gene-therapies-advisory-committee