Educators and developers are moving quickly to test whether generative AI can build students’ social-emotional skills, a domain long treated as uniquely human. Sessions at the ISTELive 26 + ASCD conference highlighted classroom and district pilots using AI chatbots to practice empathy, perspective-taking, self-awareness, social awareness, and relationship skills in low-pressure settings. Speakers including Amanda Brown, a technology implementation specialist for Montgomery County Schools in Maryland, described AI as a potential “bridge” for practicing human interaction—while emphasizing safeguards. Other educators discussed risks of cognitive offloading, where students rely on AI rather than building foundational interpersonal and critical-thinking skills. The debate is playing out across concrete tools and classroom designs. For example, instructional technology coordinator Chris Cromwell described chatbots created with School AI’s “spaces” feature that simulate disability perspectives to coach respectful question-asking and design thinking. The central issue for higher education-adjacent stakeholders is transfer: whether skills taught through AI-mediated practice can withstand reduced AI reliance in internships and early employment, and whether schools can validate learning outcomes without eroding relationship-building capacity.