We looked at the videos that had violations, and I'm 85% confident that it's caused by keywords. Words like drugs, FDA, death, murder, etc. Our headlines are specifically supposed to be sensationalistic, so we've written (or generated, rather) headlines like these: - FDA Just Opened the Door to Cheaper Drugs ๐๐งจ - FDA Pursues DEATH PENALTY in Healthcare Exec Murder ๐โ๏ธ I asked chatgpt about what keywords to avoid, and a lot of the ones we were thinking about came up in that list. So now we'll make an effort to avoid triggering keywords moving forward. Hopefully that'll help.