Meta had plenty of employees who believed that its products were addicting youth, and at its senior executive level, Meta presentations called for increasing youth engagement because youth had the lowest levels of later leaving the platform (suggesting--according to the plaintiffs--that Meta's executives knew youth users were more likely to become addicted to its use). Meta's employees and executives also appeared aware that several of the features on the site, including, for instance, photo filters, had a negative impact on many youth's body image issues, resulting in sometimes severe mental health impacts. Similarly, Meta's employees and executives appeared aware that the comment feature on its apps often led to bullying. Finally, in the New Mexico case (and perhaps others), the New Mexico AG convinced a jury that Meta's algorithms often resulted in targeting vulnerable youth for sex trafficking. These weren't merely theories plaintiffs argued; the evidence, including internal communications and the use of test accounts, was pretty damning.
Meta was likely very wise to settle these cases (the New Mexico case is separate). It had legitimate legal issues to argue on appeal, but that would only have come after a trial where Meta clearly thought it had a good chance of losing and getting hit with a big verdict.