A royal commission heard that Meta's action on hateful content fell 79% after its January 2025 policy changes, with automated removals roughly halved. For advertisers, platform-level brand safety just got weaker.
When the platform stops policing itself, the brand next to the content inherits the risk.
An Australian royal commission has heard that Meta's action on hateful conduct dropped 79% following the policy changes it made in January 2025. The company narrowed its proactive moderation, saying it would keep removing the most serious and illegal content itself but wait for user reports on less harmful material. The result was a sharp fall in enforcement. Automated removals were roughly halved, from about 35 million pieces in the six months before the change to about 17.2 million after.
Meta framed the shift as reducing enforcement errors. The commission heard a different read. Less proactive moderation means more harmful content stays up until someone reports it, and a lot of it never gets reported.
Why it matters
Every Australian advertiser buying Meta inventory should treat this as a brand safety signal. If the platform is doing less to catch harmful content before it spreads, the odds of your ad appearing beside something you would never choose go up. Platform-level safety is thinning out, which pushes the responsibility back onto advertisers and their agencies. Assuming the platform has it handled is no longer a safe assumption.
The drop in Meta's action on hateful content after its 2025 policy changes, per evidence to a royal commission
What to do about it
The platforms are stepping back from moderation, and the brands that keep buying without adjusting their own safeguards are the ones who will end up next to something they cannot defend.