As the school year kicks off and many schools globally return from break and head back to classrooms all eyes are on education and AI. As AI-powered tools make their way into classrooms, safety and appropriateness remain critical concerns. Many questions remain about these tools, alongside significant room for safe experimentation.
Top of Mind for AI and Education this week: AI in Education Trends 2025
DemandSage released their AI in Education Trends of 2025 this week. This is good synthesis of other reports and datasets in one place and it is worth a look. A few thoughts on the numbers they shared related to educator adoption of AI / educator use of AI and AI in education market dynamics in particular.
- AI tools are no longer peripheral they are becoming integral to student workflows and institutional strategies.
- This rapidly growing industry calls for clear frameworks around equity, ethical use, and pedagogical alignment. In the same breadth, we need tools that educators can use immediately to help them understand and use these things in their work.
- I am growing skeptical of just tossing money at EdTech tool development if we don’t have clear pathways to understanding actual problems that need to be solved — check out the Build4Good initiative at New America as an example of this can happen strategically and with intention towards capacity building.
- While many educators and students alike are actively embracing AI’s potential, there are still significant concerns among all users (things like data privacy, access to school work, intellectual property, etc. are pressing yet not easily or adequately addressed).
AI and Education in the UK: Building Safety Guardrails for AI in Classrooms
In their recent article Oak National Academy details its approach in developing Aila, the UK government’s first generative AI lesson-planning assistant for teachers.
To ensure that AI-generated lesson content is safe for students ages 5–16, Oak implemented four key guardrails:
- Prompt engineering to align outputs with curriculum and pedagogy.
- Input threat detection to prevent prompt attacks and malicious use.
- An independent content moderation agent (IACMA) to classify outputs as safe, requiring teacher guidance, or toxic.
- A human-in-the-loop requirement, ensuring teachers review all AI outputs before classroom use.
Oak has experienced both opportunities and challenges — such as balancing false positives with strict safety, managing edge cases like current events, and ensuring moderation adapts to real-world teacher input.
The takeaway for me: as generative AI expands in education, safety guardrails must evolve continuously, blending technical safeguards with professional judgment. This work is an early but vital step in shaping responsible AI integration in schools
As I shared at ISTE the work of Oak National Academy is a prime example of how a country can leverage AI to support significant progress in SDG 4. Key points I shared this past June include:
- Their AI tools are a national solution for open educational resources
- They model how organizations can exceed safety expectations
- Oak is continually engaging national teachers in open hackathons
- 30,000 teachers are actively using Aila, reporting they save around 3–4 hours per week in lesson planning
- Oak is sharing their code openly — something we need to see more of!
AI and Education in Asia: The Malaysia Plan Introduces the AI National Framework
In the 13th Malaysia Plan the country is prioritizing artificial intelligence and 5G to accelerate local digital transformation. The 5G Enterprise and AI City Grant was launched to fund projects with universities, public institutions, and private entities. Projects work to integrate AI and 5G to address real-world challenges with initial use cases spanning agriculture, healthcare, logistics, manufacturing, and education.
The most important thing for me in this plan is that Malaysia uses five foundational pillars that are geared toward successful implementation — something that others sometimes miss, ignore, or lay out as secondary to big ideas.
You can read more about the 13th Malaysia Plan and insights at https://opengovasia.com/malaysia-driving-digital-transformation-with-ai-and-5g/
AI and Education in Africa: New Data Shows that Africa Receives Just 0.02% of Global AI Funding in Q2 2025
This is a bit broader than education, but nonetheless it is important. Globally, AI startup funding reached $47.3 billion (making this the second-highest quarterly total). Africa remains theleast-funded region globally, with African AI startups securing only $14 million across five deals, which (if you do the math) is about 0.02% of the global total.
Inequity is ever-present and highlights the urgent need for targeted investment to unlock AI’s potential for learners across Africa
You can read more about tthe global AI funding in Q2 of 2025 at https://www.techinafrica.com/africa-receives-just-0-02-of-global-ai-funding-in-q2-2025-amid-47-3-billion-surge/.
Remember that the future of education is being shaped now and we need more voices that are informed, thoughtful, and equity-driven. Subscribe to Jody’s newsletter to get weekly updates delivered right to your in-box at https://jodybritten.medium.com/subscribe.
