Mapping the Design Space of Teachable Social Media Feed Experiences

TL;DR

This paper addresses the problem that centralized, algorithmic social feeds erode user agency by exploring teachable feed experiences through the lens of interactive machine teaching (IMT). Through a think-aloud study with 24 participants across Instagram, Mastodon, TikTok, and Twitter, the authors derive signal taxonomies (account-based and content-based) and five IMT-inspired design principles. They then instantiate these principles in three feed designs to illustrate how users could teach feeds to align with nuanced values over varying timescales. The work advances the design space for agential, personalized feeds and highlights practical paths for integrating explicit user teaching into feed curation, with ethical considerations and future research directions.

Abstract

Social media feeds are deeply personal spaces that reflect individual values and preferences. However, top-down, platform-wide content algorithms can reduce users' sense of agency and fail to account for nuanced experiences and values. Drawing on the paradigm of interactive machine teaching (IMT), an interaction framework for non-expert algorithmic adaptation, we map out a design space for teachable social media feed experiences to empower agential, personalized feed curation. To do so, we conducted a think-aloud study (N=24) featuring four social media platforms -- Instagram, Mastodon, TikTok, and Twitter -- to understand key signals users leveraged to determine the value of a post in their feed. We synthesized users' signals into taxonomies that, when combined with user interviews, inform five design principles that extend IMT into the social media setting. We finally embodied our principles into three feed designs that we present as sensitizing concepts for teachable feed experiences moving forward.

Figures (9)

• Figure 1: Examples of signals from our study. On top is a default signal template that we provide to users, and on the bottom is one completed by a participant.
• Figure 2: Areas of the Miro board. (1): signals bank with editable signal templates. (2): actions bank with sample actions and boxes for grouping content. (3): area where we uploaded participants' posts prior to the study. (4): the upper, middle, and lower feed boxes. (5): area for placing content that the participant would like to be removed from their feed.
• Figure 3: An example of a Miro board after a participant has completed the signal elicitation study. The posts themselves have been obscured by the research team to preserve the participant's privacy.
• Figure 4: Our account-based taxonomy, aggregated across Instagram, Mastodon, TikTok, and Twitter.
• Figure 5: Our content-based taxonomy, aggregated across Instagram, Mastodon, TikTok, and Twitter.