Forecasting conversational derailment is the task of predicting, as the conversation unfolds, whether it will eventually derail into personal attacks.Since forecasting models operate in an online fashion, they must decide whether to "trigger" an alert after each utterance-for example, to notify participants or a moderator that the conversation is at risk of derailing.Existing approaches make this decision solely based on the estimated likelihood of derailment given the preceding utterances, implicitly assuming that the conversation's future trajectory is fixed.As a result, they ignore the possibility of future recovery and incur an unnecessarily high rate of false positives.In this work we propose a method for decoupling the decision to trigger from the derailment likelihood estimation.Our approach is inspired by the first human baseline on this task, which shows that humans achieve dramatically lower false positive rates by selectively deferring their decision to trigger when they anticipate that tension is likely to subside.We operationalize this insight with a deferral mechanism that uses forward-looking simulations to assess whether a tense moment admits plausible paths to recovery.Incorporating this mechanism into a state-of-the-art forecasting model substantially reduces false positives without sacrificing forecasting accuracy.More broadly, this work highlights the value of treating decision making as a first-class component of forecasting systems.

Low levels of health literacy concerning Alzheimer's Disease and related dementias (ADRD) impact African American/Black communities access to appropriate ADRD care. Additionally, a legacy of mistrust in medical research due to systemic racism, has resulted in insufficient participation in ADRD clinical trials among African American/Black adults. This study explores the potential of generative AI to improve ADRD literacy and encourage participation in clinical trials among African American/Black older adults. We designed a mobile health intervention featuring AI-driven conversational agents - a chatbot and a voice assistant - specifically developed for this population. We tested the quality of the intervention using heuristics methodology adapted to the target population along with inputs from African American/ Black medical professionals and UX designers. Key findings highlight the unique needs of the African American/Black communities for culturally relevant content that is accessible to users with varying language levels and tailored to users' geographical location. Concerning the interaction, high levels of personalization and control over the interaction can promote the use of the tool, by minimizing complexity and maximizing accessibility. These findings show the novel contribution offered by our study in the domain of designing health technology with generative AI, particularly LLMS, for African American/Black communities.

During a conversation, there can come certain moments where its outcome hangs in the balance.In these pivotal moments, how one responds can put the conversation on substantially different trajectories leading to significantly different outcomes.Systems that can detect when such moments arise could assist conversationalists in domains with highly consequential outcomes, such as mental health crisis counseling.In this work, we introduce an unsupervised computational method for detecting such pivotal moments as they happen.The intuition is that a moment is pivotal if our expectation of the conversation's outcome varies widely depending on what might be said next.By applying our method to crisis counseling conversations, we first validate it by showing that it aligns with human perception-counselors take significantly longer to respond during moments detected by our method-and with the eventual conversational trajectory-which is more likely to change course at these times.We then use our framework to explore the relation between the counselor's response during pivotal moments and the eventual outcome of the session.

Contemporary highly dynamic technology and business environments, coupled with digitally savvy customers, are forcing both private and public organizations to continuously innovate and update their Information and Communication Technology (ICT) service offerings. In order to decrease the development cycles and ensure continuous delivery, many organizations are adopting new practices aimed at unifying software development and operations (i.e., DevOps). However, monolithic architectures of many large systems underpinning ICT services are severely restricting the effectiveness of such efforts. Although the latest architectural trend, microservices, is promising to solve many of the problems associated with monolithic software architectures, there is significant disagreement on when microservice architecture should be adopted, as well as how it may be best implemented. Given the limited empirical research on the topic, this paper identifies and discusses a range of opportunities and challenges associated with the adoption and implementation of microservices. The findings presented in the paper have been derived from in-depth interviews with 19 ICT architects with significant experience in large corporate systems, middleware, service oriented architectures, and, to a somewhat more limited extent, microservices.