Older adults are increasingly turning to chatbot-based mental health support, yet adoption remains limited by barriers in accessibility, privacy, security, and trust. We present a two-phase study on the co-design of MESA-Bot (Mental and Emotional Support Assistant for Older Adults), a non-diagnostic chatbot tailored to later-life needs. In Phase I, we analyzed ten leading mental health chatbots and conducted co-design sessions with N1 = 10 older adults to identify challenges and inform an accessible, privacy-aware prototype. In Phase II, we evaluated MESA-Bot with N2 = 28 older adults using semi-structured interviews and structured technical assessments. Participants emphasized transparent consent, supportive tone, and fine-grained data control. Features such as revocable consent, role-based access control, customizable data visibility, and simplified dialogue flows increased trust and usability; 86% found MESA-Bot easy to use. We offer design insights for inclusive, trustworthy mental health technologies that integrate accessibility with verifiable privacy and security protections.

Reddit is a major venue for mental-health information interaction and peer support, where privacy concerns increasingly surface in user discourse. Thus, we analyze privacy-related discussions across 14 mental-health and regulatory subreddits, comprising 10, 119 posts and 65, 385 comments collected with a custom web scraper. Using lexicon-based sentiment analysis, we quantify emotional alignment between communities via cosine similarity of sentiment distributions, observing high similarity for Bipolar and ADHD (0.877), Anxiety and Depression (0.849), and MentalHealthSupport and MentalIllness (0.989) subreddits. We also construct keyword dictionaries to tag privacy-related themes (e.g., HIPAA, GDPR) and perform temporal analysis from 2020 to 2025, finding a \(50\%\) increase in privacy discourse with intermittent regulatory spikes. A chi-square test of independence across subreddit domains indicates significant distributional differences (χ2 = 5596.67, p = 0.03, df = 50). The results characterize how privacy-oriented discussion co-varies with user sentiment in online mental-health communities.