Tomorrow (Wed), I will be presenting our @acmtochi.bsky.social paper "Purpose Mode: Reducing Distraction through Toggling Attention Capture Damaging Patterns on Social Media Web Sites" at:
- Wed, 30 Apr | 11:22 AM - 11:34 AM | Room G301
- Session link: programs.sigchi.org/chi/2025/my-...
We see the potential for Social-RAG to have broader implications for designing socially aware AI in various collaborative settings, from meeting assistants to code review bots to moderation assistants and more. 📝 Read our full paper at arxiv.org/abs/2411.02353! #RAG #HCI (9/9)
Interestingly, we also observed that groups use PaperPing very differently: in some, it enhanced engagement and encouraged sharing, while in others, it soon dominated the social space, converting channels into paper feeds. This reveals the complex social dynamics of AI in group spaces. (8/9)
Importantly, participants reported that Social-RAG enables PaperPing’s messages to achieve contextual relevance by learning from existing and ongoing user interactions (“a natural part of my day-to-day interaction while on Slack"), without requiring additional user input. (7/9)
In our 3-month field study with 18 research groups, we found that researchers valued PaperPing because its messages felt relevant, highlighting connections to past discussions, papers the group had already read, or specific members. These social signals also help groups build common ground. (6/9)
Social-RAG makes this work in four key steps:
1️⃣ Index past group conversations
2️⃣ Retrieve social signals (e.g., group interests) from the noisy interaction data
3️⃣ Feed them into a chain of prompts for generation
4️⃣ Post in-channel and gather continuous feedback
(5/9)
Based on Social-RAG, we built PaperPing, a Slack bot that recommends academic papers with tailored explanations that highlight connections to the group (e.g., referencing a previous relevant thread or mentioning specific group members who might find the paper interesting). (4/9)
Inspired by traditional RAG that retrieves factual knowledge for additional context, Social-RAG retrieves "social facts" from past group interactions to generate contextually relevant content aligned with group preferences. (3/9)
This work is done in collaboration with my amazing coauthors @xinyizhouxz.bsky.social, Lin Qiu, @josephc.bsky.social, @jbragg.bsky.social, and @axz.bsky.social (2/9)
AI agents are entering online social spaces, but often their messages feel generic or intrusive. In our #CHI25 paper, we introduce Social-RAG, a workflow that grounds AI generations in the specific group context by retrieving from the group’s interaction history. 🧵(1/9)
New paper from our lab on reporting systems! arxiv.org/abs/2306.10478 Have you ever tried to report something on social media? What do you think gets shared and who do you think sees the report? We seek to understand what ppl *think actually happens* when they report something. 1/n
This work will be presented at SOUPS next week by Leijie Wang from @socialfutureslab.bsky.social (www.usenix.org/conference/soups2023/tec... and is in collaboration with @ruotongw.bsky.social @sanketh.bsky.social and Sterling Williams-Ceci!