About Executive summary Managing risks Issues Stakeholders Transparency and remedy Annex 3. Equality and non-discrimination The right to equality and non-discrimination provides for equal protection against discrimination. As part of respecting this right, we don't allow hateful conduct on our platforms, as defined in our policy. Examples of potential inherent salient human rights risks identified in CSRA Examples of how Meta addressed the potential risks in 2025 Some languages and dialects may be more challenging to moderate than others. We introduced and open sourced Meta Omnilingual Automatic Speech Recognition (ASR), a suite of models providing ASR capabilities for more than 1,600 languages, including 500 lowresource languages. By making spoken content transcribable across previously underserved languages, this technology opens up new education opportunities and further enhances freedom of expression, empowering people to access knowledge and share content in the language they know best. Content adversely impacting equality and non-discrimination (e.g., hateful conduct) Ahead of the policy and product changes announced in January 2025, we assessed the risks of the changes and took into account relevant mitigations, such as the availability of other policies and user reports, to address this type of content. In countries experiencing conflict or crises, we may apply proactive detection by default in some policy areas, including Hateful Conduct violations. For example, we used proactive detection in Libya after deploying the Crisis Policy Protocol. 2025 Meta Human Rights Report 14