Answering “Why am I seeing this?” for Facebook and Instagram users
The problem
As a Facebook or Instagram user, you may not understand why certain content appears in your feed. In general, how AI recommender systems work is baffling to users. Plus, regulation from the Digital Services Act now requires large technology companies to provide transparency about how their algorithms work. Meta needed to meet these regulatory requirements by explaining how their core recommender systems work in a way users can understand.
The project
I created content that explains how Meta’s ranking algorithms work for common user experiences on Facebook and Instagram. Users can now see exactly what information is leveraged to rank content when they’re engaging in Feed, Reels, Stories, Marketplace, Groups, or other places on Meta’s apps. This information also now lives on Meta’s Transparency Center and is delivered in the form of what Meta calls an “AI System Card.” In a 12-month period, I created 22 System Cards for Meta’s top recommender systems. I also informed flows that get users to this information from in-app experiences.
The Opportunities
Explaining how AI works to the average user is not something most tech companies have tackled yet. The challenges and opportunities I faced on this project included:
Defining terminology. Researching and defining the correct terminology for user comprehension in a dense, complicated space.
Reaching multiple audiences. Creating content that was accessible for users, regulators, and policymakers with different levels of tech literacy.
Deep and wide XFN collaboration. Working with 22 product teams to source information, interview subject matter experts, collect stakeholder feedback, gain content approval, publish revisions, and more.
Information architecture. Organizing over 20,000 words of content into a user-friendly format and a cohesive narrative, while also considering how the content would grow and scale over time.
Content design leadership. Leading a high-stakes, content-heavy project in a strict regulatory space for a global company required me to communicate and collaborate with precision, confidence, efficiency, and empathy.
The wins
Systems Cards received wide public and user praise. Results showed that people appreciated the transparency Meta was providing and 94% of users felt the content helped them better understand and control their own product experiences within Meta apps. My team’s work was featured on media outlets like CNN, Gizmodo, and The Verge.
Regulators and policymakers also said the content met 100% of required regulatory standards, which was the primary goal behind the work.
The “What-Ifs”
If I had to do this project over, I would have loved to spend more time on user research around terminology. One of the fascinating findings of this work was learning about the wide range of tech literacy when it comes to understanding AI. Due to the time constraints of the regulatory mandate, our team had less time to spend on user research than I would have liked. I would have wanted more work around closing the gap between what users call things vs. what industry experts call them.
