AI Observatory: Waypoints & Signals – Issue 2
How can we anticipate the changes in the age of AI and help close the learning divide?
Foresight helps us make smarter decisions today, especially in the fast-evolving world of AI in education. Rather than predicting one future, it allows us to explore a range of possibilities and be proactive, preparing for changes before they happen. By considering different scenarios, we can adapt policies and strategies to stay resilient in the face of uncertainty.
Our approach to foresight is dynamic and collaborative. We use horizon scanning to spot early signs of change in AI and education, processing vast amounts of information in multiple languages and in real-time with AI and human collaboration. Through web signposting (our take on web scraping where instead of capturing vast amounts of data, we amplify existing expertise) we gather signals that help us identify gaps and where new insights are emerging, helping us stay ahead and make better-informed decisions on where to act.
Sources: 1. OECD, (2020), Horizon scanning and foresight methods. 2. Habegger, (2009), “Horizon scanning in government”. 3. UK Government Office for Science, (2022), A brief guide to futures thinking and foresight. 4. Ansoff, H. I., Managing strategic surprise by response to weak signals, (1975). 5. Cuhls et al., Models of Horizon Scanning, (2015) 6. Habegger, (2010), Strategic foresight in public policy. 7. Amanatidou et al., (2012), On concepts and methods in horizon scanning. 8. Frontier Tech Hub, (2024), Foresighting for Policy
What’s emerging?
AI supporting teachers
AI is becoming a valuable tool for teachers, especially for administrative tasks like lesson planning, assessment creation, and grading. The primary driver appears to be efficiency, with reports suggesting significant time savings; one Brazilian tool claims to reduce evaluation creation time by up to 85%. While the trend is growing, there are ethical concerns as teachers adopt AI tools while restricting student use due to fears of cheating. Platforms like Google Classroom are integrating AI features to support these tasks.
Sources: [1] Futurism, 20 Apr 2025 [2] Tempo Media, 16 Apr 2025 [3] Serambi Indonesia, 15 Apr 2025 [4] Tribun Kaltim, 15 Apr 2025 [5] Terra, 14 Apr 2025 [6] Zawya.com, 14 Apr 2025 [7] Todo Noticias, 16 Apr 2025 [8] Montevideo Portal / Montevideo COMM, 15 Apr 2025 [9] Economic Times, 15 Apr 2025
AI supporting administrators
AI is also improving school management, particularly at the governmental level. Countries like India and Brazil are integrating AI into public administration to improve decision-making, policy creation, and public service delivery.
Sources: [1] Latest Odisha News, Breaking News Today | Top Updates on Corona – OTV English, 16 Apr 2025 [2] The Times of India, 15 Apr 2025 [3] Ommcom News, 15 Apr 2025 [4] Home, 16 Apr 2025
AI tutor chatbots
AI-powered tutors and chatbots are gaining traction, offering personalised support to enhance learning. Examples include China’s Xiaoyuan AI learning machine [1][2], AI learning diagnosis tools in New Taipei [3], Agilix’s BusyBee AI teaching assistant implemented by NorthStar Academy [4], Anthropic’s Claude for Education with a ‘Learning mode’ for critical thinking [5], and Google’s Virtual Teaching Assistant pilot at the University of Michigan [6]. These tools aim to provide personalised feedback, support students with difficulties, facilitate self-directed learning, and make learning more interactive.
Sources:[1] 新浪财经, 20 Apr 2025 [2] China Daily, 16 Apr 2025 [3] 蕃新聞, 19 Apr 2025 [4] Advanced Financial Network, 16 Apr 2025 [5] Security Boulevard, 14 Apr 2025 [6] The University of Michigan, 14 Apr 2025 [7] Todo Noticias, 16 Apr 2025 [8] Economic Times, 15 Apr 2025 [9] Crunchbase News, 15 Apr 2025 [10] diginomica, 15 Apr 2025
What’s challenging?
What are the trade-offs, tensions, and challenges in the context of equitable learning in the age of AI, with a focus on Low- and middle-income countries(LMICs).
Teacher-student AI paradox
A significant tension is emerging where teachers are using AI for tasks like grading while simultaneously restricting students from using the same tools due to concerns about cheating and shortcuts. This raises ethical questions about fairness and transparency in education.
Sources: [1] Futurism, 20 Apr 2025 [2] Todo Noticias, 16 Apr 2025 [3] Montevideo Portal / Montevideo COMM, 15 Apr 2025 [4] Economic Times, 15 Apr 2025
Maintaining academic integrity
Generative AI tools are challenging traditional assessment methods, creating concerns over cheating and the authenticity of student work. This is forcing a rethink of assessment, with some questioning the viability of traditional essays and emphasising the need to understand the learning process, not just the final product.
Sources: [1] Economic Times, 15 Apr 2025 [2] Diario El Dia , 19 Apr 2025 [3] The Guardian, 18 Apr 2025 [4] Times Higher Education (THE), 14 Apr 2025 [5] Yo Soy Tu Profe, 14 Apr 2025
Exacerbating the digital divide in LMICs
While AI holds promise for democratising education, it also risks widening educational inequalities in low- and middle-income countries (LMICs). Disparities in AI access between private and public schools, as well as socio-economic barriers, highlight the need for solutions tailored to resource-limited settings.
Sources: [1] El Financiero, 15 Apr 2025 [2] Milenio.com, 16 Apr 2025 [3] Independent Newspapers Nigeria, 19 Apr 2025 [4] PressPeru, 17 Apr 2025 [5] The News International, 19 Apr 2025 [6] Advanced Financial Network, 14 Apr 2025
Teacher AI literacy gap
A gap in AI skills exists among educators, hindering effective integration. Teacher training programs are emerging to address this need in wide-ranging places from Idaho, USA to Ningbo, China, helping educators adapt their roles and develop new competencies, while growing an understanding on how AI repositions skills rather than eliminates jobs.
Sources: [1] Idaho Education News, 17 Apr 2025 [2] k.sina.com.cn, 16 Apr 2025 [3] Rosario3, 16 Apr 2025 [4] baodientu.chinhphu.vn, 20 Apr 2025 [5] Nation Newspaper, The, 16 Apr 2025
Localised AI content
There’s a need for AI tools that are culturally relevant and cater to diverse populations, especially in LMICs. The dominance of AI developed in high-income raises concerns about relevance, especially for regional and indigenous languages and marginalised communities.
Sources: [1] Zawya.com, 15 Apr 2025 [2] Zawya.com, 15 Apr 2025 [3] Forbes, 17 Apr 2025 [4] arXiv: AI, 14 Apr 2025 [5] arXiv: AI, 14 Apr 2025
Overwhelm and anxiety
The rapid pace of AI development is causing anxiety among educators and students (see some recent studies here and here). Concerns about job displacement and the overwhelming volume of new AI tools are creating uncertainty in the education sector.
Sources: [1] LatestLY, 15 Apr 2025 [2] The Ken, 15 Apr 2025 [3] educationtimes.com, 14 Apr 2025 [4] La Nacion, 18 Apr 2025 [5] Yo Soy Tu Profe, 14 Apr 2025
What’s next?
AI in education is rapidly advancing, creating new opportunities and challenges, particularly in Low and Middle-Income Countries (LMICs). What does this mean for today, tomorrow and beyond?
Upgrade: AI is already improving efficiency for teachers and administrators today, from lesson planning to grading, saving time and boosting productivity. However, ethical tensions are emerging, as teachers adopt AI while trying to navigate how students should use AI (which they already are) while balancing concerns over cheating.
Disrupt: The rise of AI-powered tutor chatbots is disrupting the learning experience of tomorrow by offering personalised support and enhancing self-directed learning. But this growth is uneven, especially in LMICs, where access to AI tools remains limited, risking a deeper digital divide.
Transform: As AI reshapes education, it raises questions around academic integrity and teacher AI literacy. There’s an urgent need to think beyond what we’ve done in the past and rethink things like assessment, promoting AI training for educators, and ensuring that AI tools are culturally relevant and inclusive for all learners, particularly in resource-limited settings.
We’d love to hear from you! What’s been shaping your thinking on AI? Drop your thoughts (and reading recommendations) over on LinkedIn.
