Video of a Speech from the Cambodian Prime Minister

Meta should update its newsworthiness allowance policy to state that content that directly incites violence is not eligible for a newsworthiness allowance, subject to existing policy exceptions. The Board will consider this recommendation implemented when Meta publishes an updated policy on newsworthy content explicitly setting out this limitation on the allowance.

Our commitment: As part of our newsworthy assessment, “[w]e remove content, even if it has some degree of newsworthiness, when leaving it up presents a risk of harm, such as physical, emotional and financial harm, or a direct threat to public safety.” We are considering ways to more precisely convey this commitment to address the recommendation.

Considerations: Under our existing Violence and Incitement policy, we do not allow content that includes language that may incite or facilitate serious violence. In rare instances, if we have determined that content might violate one of our policies, we may allow content if it is newsworthy and keeping it visible is in the public interest.

To make this determination, we balance the public interest value of the content against the risk of harm. We consider a number of factors, including whether the content surfaces an imminent threat to public health or safety or gives voice to a perspective currently being debated as part of a political process, as well as country-specific circumstances, the nature of the speech, and the political structure of the country (including the existence of a free press). When applying our newsworthy balancing test, we remove content "when leaving it up presents a risk of harm, such as physical, emotional and financial harm, or a direct threat to public safety."

We are reviewing the clarity of our newsworthiness language and, if we change that guidance, will share the update in our Transparency Center.

Meta should immediately suspend the official Facebook Page and Instagram account of Cambodian Prime Minister Hun Sen for a period of at least six months under Meta's policy on restricting accounts of public figures during civil unrest. The Board will consider this recommendation implemented when Meta suspends the accounts and publicly announces that it has done so.

Our commitment: Upon assessing Hun Sen’s Facebook Page and Instagram account, we determined that suspending those accounts outside our regular enforcement framework would not be consistent with our policies, including our protocol on restricting accounts of public figures during civil unrest.

Considerations: We have removed the content that was the subject of this case and, consistent with our policies, applied appropriate account-level penalties associated with that action. There is not currently any basis to suspend Hun Sen’s account under our policies.

As discussed in our response to Recommendation #1, we determined that suspending Hun Sen’s accounts is not appropriate under our protocol on restricting accounts of public figures during periods of civil unrest. In making this determination, we relied on our Crisis Policy Protocol (CPP) to assess on and off-platform risks of imminent harm at the time Hun Sen made the post in question, to determine whether conditions in Cambodia constituted a crisis. Based on an assessment of the framework criteria contained in the protocol - including levels of ongoing civil unrest/violence, state-imposed restrictions on media, and any significant reductions of human rights as compared to the baseline conditions in the country - we determined that the situation in Cambodia did not meet the entry criteria threshold for crisis designation under the Crisis Policy Protocol.

Hun Sen’s accounts will continue to be subject to our generally applicable penalty system. Our Community Standards apply to all users all around the world. When a user violates these standards, they may also incur a strike to their account. We notify users when this happens and, if they continue to violate our policies, we may restrict certain account features as detailed in our Transparency Center.

We will have no further updates on this recommendation.

Meta should update its review prioritization systems to ensure that content from heads of state and senior members of government that potentially violated the Violence and Incitement policy is consistently prioritized for immediate human review. The Board will consider this recommendation implemented when Meta discloses details on the changes to its review-ranking systems and demonstrates how those changes would have ensured review for this and similar content from heads of state and senior members of government.

Our commitment: We will assess the technical feasibility and weigh the tradeoffs of ensuring that content posted by heads of state that is reported for potentially violating our Violence and Incitement policy is more consistently prioritized for human review relative to other high-severity content.

Considerations: Our systems generally surface potentially violating content from heads of state and senior members of government for human review in several ways, including through our normal proactive and reactive detection systems as well as via existing escalation pathways and our cross-check program. We are assessing the feasibility of adding additional signals to classifiers as another avenue to ensure we surface and review content that potentially violates our Violence and Incitement policy posted by this type of public figure.

We have enrolled the vast majority of heads of state and senior members of government into our Early Response Secondary Review (ERSR) program, which we have expanded upon in our responses and updates to the Cross-Check PAO. Our systems prioritize content according to severity of violation in alignment with our Community Standards. Where we encounter potentially violating content which can be exposed to large audiences, our early response program has ranking mechanisms to ensure that we review the content within a designated time frame and take the necessary action. As they operate within the ERSR program, reported content from heads of state consistently qualify for human review when flagged for any violation, including potential Violence and Incitement.

Where our ERSR program and detection systems have not captured potentially violating content for human review by default, we have escalation channels and appeals processes to address anomalies. However, as cross-check is primarily intended to prevent false positive violations rather than identify false negatives on potentially violating content, we are assessing the feasibility of adding additional signals to classifiers that determine the severity of content to ensure that this type of content is more effectively prioritized for human review.

While we believe our current systems, including cross-check and escalations based on regional markets context, human rights considerations, and trusted flaggers inputs, ensure that the vast majority of potential violence and incitement from heads of state or senior government officials is currently prioritized for efficient human review, we are also considering opportunities to capture even more content that may fall into this category. We are considering opportunities to adjust the criteria for entry into high-severity review queues, including additional signals that specify that the content was posted by a head of state and that it likely has a high viral velocity.

Meta should implement product and/or operational guideline changes that allow more accurate review of long-form video (e.g. use of algorithms for predicting the timestamp of violation, ensuring proportional review time with length of the video, allowing videos to run 1.5 times or 2 times faster). The Board will consider this implemented when Meta shares its new long-form video moderation procedures with the Board, including metrics for showing improvements in review accuracy for long-form videos.

Our commitment: Our long-form video review optimization processes and protocols remain a priority across our operational teams. We already have a number of established processes geared towards improving video review accuracy and we will continue to iterate on new improvements. We will also evaluate metrics to demonstrate ongoing improvements in review accuracy for long-form videos to the public and to the Board.

Considerations: We currently have operational guidelines in place that govern long form video review. Our protocol seeks to ensure accurate review of long form video while optimizing for the efficiency and wellbeing of our reviewers. Our guidelines include a designated review protocol, explicit minimum review standards including duration as well as video indicators such as violation and reporting timestamps and speed adjustments. Our guidelines also ensure that we apply progressive review on long form video including the assessment of title, description, caption, thumbnails, transcript and comments attached to the video under review.

We have developed new categorical guidance pertaining to live videos and provided reviewers with the corresponding action hierarchy for such instances, still optimizing for both reviewer precision and accuracy with widgets such as “watch this” indicators for potentially violating content as well as warning signs and interstitials. Additionally, we have a designated team that engages with live video reviewers on a recurring basis to surface protocol and tooling improvements. All our video review processes include video specific resiliency features such as muting video, adjustable skip lengths and blurring. We continue to innovate our efforts on video review with solutions such as AI thumbnail generation to flag potentially violating content, increase the accuracy of the existing thumbnails and support reviewer efficiency.

We have rolled out efforts to detect looping videos which are long-form types used by repeat offenders on our platform. We have completed automation to roll this out on Facebook and will scale to Instagram in the second half of 2023.

We continue to improve our products and protocols for long-form and live video review—this remains an ongoing priority area across our enforcement teams. We will also continue to assess the most effective way to use metrics to publicly demonstrate review accuracy for long form videos and share these in future updates.

In the case of Prime Minister Hun Sen, and in all account-level actions against heads of state and senior members of government, Meta should publicly reveal the extent of the action and the reasoning behind its decision. The Board will consider this recommendation implemented when Meta discloses this information for Hun Sen, and commits to doing so for future enforcements against all heads of state and senior members of government.

Our commitment: We may publicly explain the rationale behind future account-level enforcement actions against heads of state or senior members of government, but will do so on a case-by-case basis, balancing transparency and security considerations.

Considerations: While we have shared details about enforcement actions on the accounts belonging to Hun Sen in this case, and on the accounts of former U.S. President Trump, there may be circumstances where privacy and security considerations weigh against Meta publicly sharing details about actions taken on an account. Given these considerations, we cannot commit to always sharing all details about account restrictions on accounts of heads of state or senior members of governments.

Having integrated the Board’s guidance in the Trump Decision into our approach for account-level suspension of heads of state, we will continue to aim for transparency around this type of enforcement on a case-by-case basis. We consider this recommendation implemented in part and will have no further updates.