According to Ethan Mollick on X, a preregistered study of 968 participants found almost no correlation between feeling empathic and communicating empathy, but a single practice session with an AI coach measurably improved empathy communication skills. As reported by the study authors on arXiv, participants who practiced responses with an AI coach showed statistically significant gains on validated empathy communication scales after one session, indicating rapid skill transfer in hard-to-teach social behaviors. According to the arXiv paper, the intervention used structured practice with feedback to close the gap between internal empathic concern and externally observable empathic communication, highlighting immediate applications for customer support, healthcare triage, and manager training programs. As reported by arXiv, the preregistered design and large sample size strengthen external validity for enterprise learning and development teams evaluating AI-enabled soft-skills training.

According to Google Gemini on X (@GeminiApp), users can manage which apps connect to Gemini, provide thumbs up or down feedback, correct outputs inline, and toggle Personal Intelligence off per chat to refine personalization. As reported by the Google Gemini post, these user controls directly influence recommendation quality and privacy, giving enterprises clearer governance over data flows when deploying Gemini-powered assistants. According to the Gemini announcement, per-chat opt outs and explicit feedback loops enable safer experimentation, faster model alignment to user preferences, and reduced risk of oversharing context across workflows, which can lower support costs and improve CSAT in customer-facing AI deployments.