Understanding Help Seeking for Digital Privacy, Safety, and Security

TL;DR

This study investigates how users seek help for digital privacy, safety, and security on Reddit at scale, revealing a rapidly growing need and a broad, context-rich set of support requirements. By combining qualitative coding with large-scale LLM fine-tuning, the authors identify 3 million help-seeking posts across 5,401 subreddits, annotating each with nine topics and dominant emotions, achieving high precision and recall. The work shows that help-seeking is complex, context-dependent, and unevenly distributed across topics and communities, suggesting tailored, topic- and platform-specific resources. The findings inform the design of LLM-assisted help-giving agents, topic-specific guides, and community-organization strategies, and they advocate for ecosystem-level collaboration to better meet users' sensemaking, guidance, and therapeutic needs. The authors also release their URLs and annotations to support further research and benchmarking in digital privacy, safety, and security resources.

Abstract

The complexity of navigating digital privacy, safety, and security threats often falls directly on users. This leads to users seeking help from family and peers, platforms and advice guides, dedicated communities, and even large language models (LLMs). As a precursor to improving resources across this ecosystem, our community needs to understand what help seeking looks like in the wild. To that end, we blend qualitative coding with LLM fine-tuning to sift through over one billion Reddit posts from the last four years to identify where and for what users seek digital privacy, safety, or security help. We isolate three million relevant posts with 93% precision and recall and automatically annotate each with the topics discussed (e.g., security tools, privacy configurations, scams, account compromise, content moderation, and more). We use this dataset to understand the scope and scale of help seeking, the communities that provide help, and the types of help sought. Our work informs the development of better resources for users (e.g., user guides or LLM help-giving agents) while underscoring the inherent challenges of supporting users through complex combinations of threats, platforms, mitigations, context, and emotions.

Figures (9)

• Figure 1: Automated pipeline for identifying posts seeking help for digital privacy, safety, or security. Human experts reviewed thousands of Reddit posts to define help seeking and to identify nine distinct help-seeking topics. These same experts annotated a golden dataset for fine-tuning which allowed an LLM to generalize and apply these qualitative definitions at-scale, in addition to isolating the emotion of the post.
• Figure 2: Volume of monthly help-seeking activity on Reddit for digital privacy, safety, and security concerns. (Note truncated y-axis.)
• Figure 3: Temporal trends in help-seeking topics. Each line represents the ratio of posts that contained a topic versus all help-seeking posts, per month.
• Figure 4: Heatmap showing Pearson correlations between each pair of help-seeking topics.
• Figure 5: Top dominant emotions identified in posts for each of the nine help-seeking topics.