Education systems should not delay efforts to develop guidance on the use of AI, as students and teachers already have independent access, and many existing technologies embed AI into their systems (e.g., search engines, email applications, schedulers, assistive technologies, etc.). Ensuring that AI use complies with existing security and privacy policies, providing guidance to students and staff on topics such as the opportunities and risks of AI, and clarifying responsible and prohibited (especially uses that require human review and those related to academic integrity and grading) will shape AI educational ecosystems. Establishing these parameters and guardrails depends on basic AI literacy, so that guidance and policies will mitigate risks without thwarting organizational learning or stifling improvement and transformation. At the same time, they also lay the foundations for advancing an AI literacy that aligns with student safety, privacy, and learning.

In parallel with guidance and policies that address immediate concerns, education systems must prioritize professional development for all staff and bring together individual educators' experiences with AI to document successes, identify gaps, and build collective organizational knowledge and capacity. A systemwide approach would go beyond instructional issues to include operational considerations, such as evaluating AI tools already in use, creating selection criteria for future evaluations, and understanding how AI is used across departments. It also leads to more equitable AI integration across classrooms and prevents inequities from emerging when innovation diffuses haphazardly.

Thoughtful guidance and a commitment to ongoing organizational learning are essential to navigating the uncertain and evolving role of AI in education. While AI presents promising possibilities, such as supporting personalized learning, aiding formative feedback, and assisting with certain administrative tasks, these outcomes are not guaranteed and should be approached with caution. Any perceived efficiency gains, for example, must be weighed against the time and resources needed for professional development, implementation, and ongoing evaluation. Incorporating AI into education creates a new context for reexamining long-standing priorities, such as equity, access, and instructional quality, but it does not inherently advance reform.

The extent to which AI use results in productive change or exacerbates existing challenges depends on how it is introduced, governed, and integrated. Strong guidance, continuous organizational learning, and a systems approach to improvement work together to help educational institutions make informed, adaptive decisions. Rather than assuming a linear path toward transformation, this approach embraces an ongoing cycle of thoughtful introduction, evaluation, and adjustment, recognizing that both opportunities and concerns must be addressed continuously and in tandem.

AI Consultancy Protocol (EDSAFE AI Alliance): A step-by-step process for assessing readiness and building capacity for ethical AI implementation.