Facebook Uses CLARA to Improve Community Standards Enforcement

Facebook has seen the ability of AI to help them fight the misuse of data. Among fake Facebook accounts, 99.5% are taken down before users report them. AI is also used when the platform adds scam warnings on Messenger last 21 May 2020. It also used AI as it updated the news feed algorithm on 30  June 2020. Recently, Facebook uses CLARA to improve its community standards enforcement.

Facebook

CLARA stands for Confidence of Labels and Raters. It’s a machine-learning tool that estimates the uncertainty of human decisions. It’s designed to help raters make the right decision on enforcing Facebook’ Community Standards. Facebook adds such an AI tool to prevent human biases. Thus, it ensures more accurate enforcement of the platform’s policies. On Facebook’s side, using CLARA reduces the need for human moderation on the said category. This enables them to maximize manpower.

Facebook uses CLARA to improve its community standards enforcement effective 21 August 2020.

Implications for Marketers:

Facebook makes a smart move using CLARA to provide unbiased policy enforcement. It enables the platform to focus on key areas needing human decisions. For marketers, this is an assurance of smooth policy imposition on Facebook. Moreover, it can help them better understand audience preferences. 

Reference: https://research.fb.com/blog/2020/08/improving-the-accuracy-of-community-standards-enforcement-by-certainty-estimation-of-human-decisions/