Meta launches third-party verification for feed powered by AI

The new inventory controls by Meta for Facebook and Instagram Feeds are now rolling out to advertisers in English and Spanish-speaking markets. The third-party verification solution for Facebook Feed is also now available through Zefr and additional Meta Business Partners will onboard in the coming months.

Facebook's new updates highlight the ongoing collaboration with industry partners and the inclusion of AI into advertisers' functionalities such as brand suitability controls on Meta & third-party verification.

AI-powered Controls

Meta built a multi-stage AI review system to classify content for advertisers to ensure brand suitability controls are in place. These models will complement our existing technology, which aims to identify content that violates or potentially violates Community Standards and Guidelines.

This system learns to classify content in Facebook and Instagram Feeds – not only text but also video and images – to determine if it meets monetization policies; if it does not, the content is not eligible to have ads appear above or below it. When content is found to be eligible for ad adjacency, the models assign it to a suitability category.

The social network has collaborated with industry partners such as Global Alliance for Responsible Media (GARM), an industry trade organization created to ensure that harmful content is not monetized. Meta has developed controls that align with GARM’s Suitability Framework, which defines high, medium, and low-risk content. Advertisers can now choose from the following three settings to control the type of monetizable content that can appear above and below an ad:

Also Read: Pinterest adds new tools & features to enhance shopping experience

• Expanded inventory: This is the default setting and shows ads next to content that a) adheres to our Community Standards and b) meets our monetization eligibility criteria.
• Moderate inventory: For advertisers who may want to take a moderately conservative approach, this filter excludes content that may be considered high risk, in alignment with the GARM Brand Suitability Framework.
• Limited inventory: For advertisers who want to take the most conservative approach, this filter excludes content that may be considered both high and medium risk, in alignment with the GARM Suitability Framework.

Meta will expand these controls to support additional languages and make them available to advertisers in more countries, and also start testing them on additional placements including Reels, Stories, Video Feeds, and other surfaces across Facebook and Instagram, to improve and enhance this technology.

Third-party Brand Suitability Verification

The company has also been working with Zefr to test and build an independent AI-powered solution to report the context in which ads appear on Facebook Feed. In early testing, it was found through third-party verification with Zefr, that less than one percent of the content on Facebook Feed falls into the high-risk GARM suitability category.

Zefr’s AI product assesses video, image, text, and audio to label Feed content based on the GARM suitability standards. The solution allows advertisers to measure, verify and understand the suitability of content near their ads to help them make informed decisions in order to reach their marketing goals. Meta will be rolling out this verification and measurement solution to additional badged Meta Business Partners this year.