The take up of AI in education in the last year in a global setting

1. Strategic Overview: The 2026 Landscape

As of January 2026, tracking AI adoption in global science education has moved from experimental curiosity to a strategic necessity. Dr. Glenys Hart—International Education Consultant and expert physics teacher/Inspector—assessed this landscape by auditing audience knowledge. This methodology tracks the "take-up" of AI through a transition from personal ChatGPT use to professional integration of VR and robotics. Crucially, the data indicates that AI is not replacing the teacher or the child; rather, it is augmenting traditional roles. This progression requires a framework that aligns these technologies with established pedagogical goals.

2. The Tripartite Framework: Language, Images, and Robots

The session categorizes AI integration into three domains—Language, Images, and Robots—matching distinct learning styles with specific technical affordances.

• Language-based AI: Strategists move beyond basic prompts to specialised "Pro Tips." Claude is prioritized for its ability to summarize 200k+ tokens for deep reading, while Perplexity is the standard when accuracy and verified citations matter.
• Visual AI: Breaking geographical barriers, platforms like NASA and the Faulkes Telescope Project enable students to engage in the remote control of professional equipment in Hawaii and Australia, rather than just viewing static data.
• Robotics: Social robots (Pepper, NAO, Saya) utilize a "human-in-the-loop" model. These tools facilitate scripted concept teaching while ensuring educators retain primary responsibility for safeguarding and learning outcomes.

3. Cross-Disciplinary Case Studies and Institutional Research

Strategic adoption relies on "approved services" like Google NotebookLM, which—as seen at Heathfield Community College and the University of Cambridge—provides a secure environment for citation-based research.

4. Leadership, Governance, and Evaluative Frameworks

Effective AI implementation requires an organizational shift toward ethical oversight. Inspectors now seek technical evidence, such as Data Protection Impact Assessments (DPIAs) and Equality Impact Assessments, to satisfy "Ofsted-style" scrutiny:

• Strategic Alignment: Does AI integration reflect core institutional values?
• Risk Mitigation: Are bias, privacy, and safeguarding addressed through DPIAs?
• Workload vs. Impact: Does the tool provide genuine efficiency gains for staff?

Furthermore, AI promotes inclusion through vendors like No Isolation, whose robots represent absent friends for students unable to attend physically. Ultimately, the session concludes that science pedagogy in the AI era must be rooted in evidence-based decision-making and continuous improvement.