Meta’s Policy Changes Halve Erroneous Content Removals Without Increasing Harmful Exposure

Meta has implemented significant changes to its content moderation policies on Facebook and Instagram, focusing more on supporting ‘free expression’ and reducing enforcement errors. Announced in January 2025, these adjustments included relaxing certain rules and relying less on automated moderation tools, particularly for less severe violations. As a result, Meta’s latest Community Standards Enforcement Report reveals a substantial decrease in content removals: 1.6 billion items were removed globally from January to March 2025, compared to nearly 2.4 billion in the previous quarter. This represents a one-third drop in takedowns, with notable reductions in removals for spam (down 50%), child endangerment (down 36%), and hateful conduct (down 29%). The only category with increased removals was suicide and self-harm content. Meta attributes this decline largely to improved accuracy and fewer enforcement mistakes, leading to fewer appeals and restored posts. Despite concerns from some employees about the potential for increased harmful content, Meta’s internal data suggests user exposure to hateful content actually decreased slightly, averaging one to two instances per 10,000 posts viewed. The company’s platforms also continued to grow, with 3.43 billion users in March 2025. The changes, however, have sparked debate, particularly regarding the relaxation of language rules that some human rights groups find problematic.