The increasing presence of generative artificial intelligence (GenAI) in schools raises questions regarding how teachers interpret this technology and integrate it into instructional and assessment practices. This study examines variations among teachers by identifying profiles using cluster analysis based on attitudes toward GenAI, perceived pedagogical benefits, concerns regarding students’ use of GenAI, AI-related anxieties, and pedagogical orientations. This analysis is based on survey data collected from January to March 2025, involving 501 Israeli teachers from various educational sectors.

Findings revealed three profiles. The Closed and Skeptical profile, the Open but Anxious profile, and the Open and Confident profile. While teachers in the Open and Confident profile reported active use of GenAI alongside pedagogical change, and those in the Closed and Skeptical profile expressed resistance and limited openness to the technology, the Open but Anxious group represented the profile with the greatest potential for change, as it combined openness toward GenAI with hesitation and concern regarding use.

Additional analysis identified three sub-profiles: Implementing Despite Concerns, Cautious Observers, and Anxiety-Constrained Teachers.

Findings indicate that AI-related anxiety functions as a differentiating factor across teachers’ profiles. Similar levels of anxiety were associated with fundamentally different pedagogical patterns, depending on how anxiety intersected with the perceived pedagogical value of GenAI. Sometimes, anxiety was linked to limited use and minimal pedagogical change, whereas in others, it coexisted with extensive use and substantial instructional and assessment-related change.

The results underscore the need to move beyond one-dimensional adoption indicators and emphasize the importance of differentiated professional development and implementation strategies aligned with distinct teacher profiles, thereby increasing the likelihood of effective and sustainable GenAI integration in schools.