Public Views Teaching as Coaching, Not an AI Task

More than 3,000 people across the United States, England and Australia overwhelmingly view teaching as a relationship-driven practice akin to coaching, largely ignoring the artificial intelligence tools dominating current education policy debates.

A survey conducted between May 2025 and March 2026 asked respondents in all three countries to compare teaching to another occupation or activity. Coaching emerged as the top answer everywhere, selected by 30% of U.S. participants and 23% of those in both England and Australia. The results challenge a prevailing narrative in education technology circles that frames instruction primarily as information delivery ripe for automation.

While the consensus on coaching crossed borders, the secondary comparisons revealed distinct cultural frameworks. Respondents in England and Australia frequently aligned teaching with credentialed institutional professions such as nursing, social work and university lecturing. A 39-year-old Australian man explicitly linked teachers to nurses, noting both care for people at their most vulnerable while building essential life skills.

Americans took a different path. U.S. respondents were far more likely to equate teaching with parenting or religious leadership, invoking figures such as priests, pastors and rabbis. This divergence reflects deeper structural differences in how each society views the state’s role in child-rearing. In the U.S., where local control of schools is deeply entrenched and parental rights movements have surged in recent years, the boundary between family caregiving and classroom instruction remains highly porous. England and Australia, which operate under more centralized national curriculum frameworks, tend to conceptualize educators closer to clinical or academic professionals.

The emphasis on interpersonal connection also transcends domestic political divides. Previous research examining U.S. attitudes found that Republicans and Democrats share remarkably similar definitions of effective teaching, despite intense partisan battles over school funding, library books and curriculum standards. The international data suggests this alignment extends well beyond American borders.

Perhaps the most striking finding is what the public omitted. Artificial intelligence has consumed educational discourse throughout 2025 and 2026. School districts nationwide are currently drafting policies on generative AI use, and ed-tech companies are aggressively marketing automated tutoring platforms. Yet out of thousands of responses, AI was mentioned only a handful of times and never served as the basis for comparing teaching to another field.

This public blind spot contrasts sharply with expert priorities. An informal September 2026 survey by The New York Times gathered 30 proposals from 37 education researchers aimed at fixing schools. Their recommendations focused heavily on systemic interventions: increasing per-pupil funding, banning cellphones and restricting artificial intelligence in classrooms. None of these policy-level concerns mirrored the relational focus of the general public.

Historically, the tension between teaching as knowledge transmission and teaching as mentorship dates back to early 20th-century pedagogical debates. Progressive education reformers like John Dewey argued for experiential, student-centered learning, while traditionalists favored direct instruction. Modern cognitive science has largely validated the progressive view, demonstrating that students retain information significantly better when they feel psychologically safe and personally supported by an instructor.

The disconnect between policymakers prioritizing technological guardrails and citizens prioritizing human connection presents a tangible challenge for school administrators. Districts investing heavily in AI-driven instructional software may find community resistance if those tools are perceived as replacing the mentoring dynamic parents and voters expect.

Gustavo E. Fischman, Margarita Pivovarova and Eric Haas led the research through Arizona State University and California State University, East Bay.