AI in education: what’s next?

Dr Martin Compton, AI and Innovation in Education Lead

Huinan Zeng, AI manager with a PhD in Psycholinguistics

In the previous step, you looked at ways in which curricula may need to change. In this step, you will begin by thinking about imminent impacts and tensions across education. What does the future look like?

The pace of change: rapid tech versus slow institutions

Universities and schools are often slow to adopt new practices, constrained by established structures, limited resources, and the need for consensus. In contrast, the tech landscape is evolving incredibly fast, with AI ‘agents’ having the potential to challenge traditional approaches to teaching, learning, and administration. This tension underscores the need for a phased and thoughtful approach to AI adoption.

Some predictions for AI in education

• Pedagogic and curricular evolution
  • Institutions must rethink both how they teach and what they teach. For example, medical degree curricula will need to reflect advances such as AI image diagnostics to prepare students for cutting-edge applications in their fields.
  • In other disciplines, from business to the humanities, the integration of AI literacy into general education will become essential to ensure graduates are equipped for the evolving job market.
• Seamless integration of AI into everyday tools
  • AI functionalities will become increasingly embedded in tools we already use, such as email, word processors, and spreadsheets. These AI-supported features will work invisibly in the background, streamlining workflows and reducing cognitive load for educators and students alike. Indeed, we are increasingly seeing AI in mobile devices and search engines. A good question to ask ourselves is: what, if anything, is AI adding or enabling?
  • AI tools integrated into a word processor might offer real-time feedback on writing style, grammar, and even content relevance, without requiring additional input from the user. Microsoft 365 Copilot, introduced in March 2023, is already embedded into all Microsoft 365 apps, including Word, Excel, PowerPoint, Outlook, and Teams. Its capabilities have also been continuously updated since its release. There is no reason to assume a cessation or slowing of such processes as AI becomes more readily available.
• The blurring line between search engines and AI chatbots
  • Search engines and large language model (LLM) chatbots are likely to fuse into a single phenomenon, redefining how we access and process information. This shift will create more dynamic and context-aware tools for research and learning, capable of providing nuanced answers and synthesising information across multiple sources. At the same time this will likely (in the short term at least) exacerbate existing energy cost issues, as well as data reliability and provenance concerns.
• Advances in voice and communication tools
  • Dialogic tools, such as the voice interfaces already available on mobile platforms like ChatGPT and Google Gemini, will continue to improve. These tools will offer smoother, more accurate conversational experiences, enhancing accessibility and engagement. Changes to generative tool interfaces will enable increased diversity of use with minimal technical upskilling required. NotebookLM’s experimental interface is one example of how the already familiar LLM chatbot, typified by ChatGPT, might be repackaged to focus on different functionalities (in this case cross-source synthesis, analysis, and comparison).
  • Coupled with advances in transcription and translation, these technologies have the potential to revolutionise communication, enabling seamless learning and research across languages and fostering global collaboration.

Emerging tensions: tools versus agents

The distinction between tools and agents will become increasingly important:

• Tools require user input and serve specific, reactive purposes (eg a spellchecker or grade calculator).
• Agents operate independently, supporting broader, dynamic goals by anticipating needs and taking initiative (eg a scheduling assistant that resolves calendar conflicts proactively).

Custom GPTs illustrate this evolution, straddling the line between tools and agents. While they can function as simple chatbots, with the right integrations they can evolve into intelligent agents capable of executing complex tasks, such as managing research workflows or providing tailored academic support.

Opportunities and challenges for education

AI agents and tools can play transformative roles in education, but their adoption comes with unique challenges. Many of these opportunities and challenges have been raised throughout the course, but here is a summary to remind you:

Technological aspect

• Stability and capability: while AI capabilities expand rapidly, their reliability varies significantly. In education, stability must take precedence over novelty.
• Measuring success: AI systems involve interconnected tasks, making success metrics complex. Clear impact analysis and thoughtful feedback collection are essential for guiding development.
• Energy and resources: AI solutions require significant computational resources, raising sustainability concerns.
• Personalisation versus privacy: AI can analyse student performance data to proactively provide insights or suggest actions, but this raises privacy concerns about machine processing of personal information.
• Administrative integration: AI can streamline processes, but it requires balanced human oversight and transparent systems.

Human aspects

• Policy and guidance: rapid AI advancement requires continuous updates to institutional policies and regulations, particularly regarding copyright and responsible use.
• AI literacy: while younger generations adapt more naturally to AI, many educators must actively learn new methods and unlearn traditional approaches.
• Job security concerns: staff may resist AI adoption despite its benefits, fearing eventual role displacement.
• Accessibility implementation: AI tools offer support for diverse learners but require careful implementation and staff training.
• Student engagement: AI monitoring can improve student engagement and retention through early intervention, but it must balance support with student privacy and autonomy.

The road ahead

The integration of AI into education is inevitable, but its success will depend on the ability of institutions to adapt. This adaptation must encompass both technological and cultural shifts:

• Technological integration: institutions should focus on embedding AI seamlessly into existing systems and workflows, ensuring its benefits are accessible to all.
• Cultural change: educators and administrators need to embrace the possibilities of AI while addressing ethical concerns, data privacy, and the risk of over-reliance on technology.
• Global collaboration: enhanced voice, transcription, and translation tools can break down language barriers, fostering international partnerships and broadening access to knowledge.

As we look ahead, the question is not whether AI will transform education, but how quickly and effectively it can be harnessed and managed (or, indeed, whether it can) to meet the needs of students, educators, and society at large. The balance between innovation and tradition will be key to navigating this exciting and challenging frontier. Where possible, we should engage students in the design of AI tools for education. According to the Digital Education Council Global AI Student Survey 2024 (with 3,839 responses from 16 countries), 71% of students want to be more involved in AI decision-making at their universities. However, only 34% believe their university is actively seeking their feedback [1].

And where are the even younger voices? ‘The Children’s Manifesto for the Future of AI’ is a worthwhile resource both to elevate young people’s concerns and as a vehicle for continuing these discussions in classrooms. Whatever comes next, we need to ensure it is not just the big tech companies driving the narratives. Teachers and students need their voices heard too. It is their future that is being impacted, after all.

Sharing your research

How would you like to contribute something about your exploration or research in the AI space to a growing repositiory? We at King’s are partnered with the University of Cambridge supporting the ‘Camtree’ project: Cambridge Teacher Research Exchange. This initiative promotes and supports the sharing of close-to-practice research carried out by teachers working in any educational setting.

Could your practice-based research become part of this repository? Find further details in the attached PDF under Downloads at the end of this page.

Now that you have completed this step you have seen some of the advances taking shape as we assembled this course, as well as some likely tensions. The final steps will give you an opportunity to review your understanding and reflect on your learning in Week 3 and across the whole course.

References

1. Digital Education Council. Digital Education Council Global AI Student Survey 2024 [Internet]. 2024 [cited 2025 Mar 19]. Available from: https://www.digitaleducationcouncil.com/post/digital-education-council-global-ai-student-survey-2024

Join the conversation

What do you think the future holds for education? How much might educators’ and young people’s voices be heard and taken into account?