What if mobile learning had the intelligence and context it lacked 25 years ago? This piece revisits the rise and fall of early mobile learning projects and considers how the convergence of artificial intelligence, contextual mobile data and educational expertise could support more responsive and personalised learning today.

Bringing Mobile Learning Back with AI, Context and Expertise

Author: Prof John Traxler, UNESCO Chair, Commonwealth of Learning Chair and Academic Director of the Avallain Lab

St. Gallen, July 28, 2025 – Around 25 years ago, many members of the European edtech research community, myself included, were engaged in projects, pilots and prototypes exploring what was then known as ‘mobile learning’. This roughly and obviously referred to learning with mobile phones, likely 3G, nearing the dawn of the smartphone era. Learners could already access all types of learning available on networked desktops in their colleges and universities, but they were now freed from their desktops. The excitement, however, was around all the additional possibilities. 

One of these was ‘contextual learning,’ meaning learning that responded to the learner’s context. Mobile phones knew where they were, where they had been and what they had been doing¹. These devices could capture images, video and sound of their context, including both the user and their surroundings. This meant they could also understand and know their user, the learner. 

So, to provide some examples:

• Walking around art galleries like the Uffizi and heritage sites like Nottingham Castle, learners with their mobile phones could stop at a painting randomly and receive a range of background information, including audio, video and images. The longer they stayed, the more they would receive. Based on other paintings they had lingered at, they could get suggestions, explanations and perspectives on what else they might like and where else they could go.

• Augmented reality on mobile phones meant that learners standing in Berlin using their mobile phone as a camera viewfinder could see the Brandenburg Gate, but with the now-gone Berlin Wall interposed perfectly realistically as they walked up to and around it. Similarly, they could see Rembrandt’s house in Amsterdam. Learners could also walk across the English Lake District and see bygone landforms and glaciers, or engage in London murder mysteries, looking at evidence and hearing witnesses at various locations.

• Recommender systems on mobile phones analysed learners’ behaviours, achievements and locations to suggest the learning activity that would suit them best based on their history and context. These recommendations could be linked to assignments, resources and colleagues on their university LMS, providing guidance and practical advice. For example, in a Canadian project, there are specific applications in tourism.

• Using a system like Molly Oxford on their mobile phones, learners could be guided to the nearest available loan copy of a library book they wanted. They could also be given suggestions based on public transport, wheelchair accessible footpaths and library opening hours.

• Trainee professionals, such as physiotherapists or veterinary nurses, in various projects across Yorkshire, could be assessed while carrying out a healthcare procedure in ‘real-life’ practice. Their mobile phones would capture the necessary validation and contextual data to ensure a trustworthy process.

• Some early experiments, with Bluetooth and other forms of NFC (near-field communication), allowed passers-by or students to pick up comments or images hanging in discrete locations, such as a subway or corridor on a university campus, serving as sign-posting or street art. 

These pilots and projects implemented situated², authentic³ and personalised⁴ learning as aspects of contextual learning, and espoused⁵ the principles of constructivism⁶ and social constructivism⁷. This was only possible as far as the contemporary resources and technologies permitted. They did not, however, encourage or allow content to be created, commented on, or contributed to by learners, only consumed by them. Also, they usually only engaged with learners on an individual basis, not supporting interaction or communication among learners, even those learning the same thing, at the same place and at the same time.

So what went wrong? Why aren’t such systems widespread across communities, galleries, cultural spaces, universities and colleges any more? And how have things changed? Could we do better now?

The Downfall of Mobile Learning: What Went Wrong?

Mobile phone ownership was not widespread two decades ago, and popular mobile phones were not as powerful as they are today. The ‘apps economy’⁸ had not taken off. This meant that projects and pilots had to develop all software systems from scratch and get them to interoperate⁹. They also had to fund and provide the necessary mobile phones for the few learners involved¹⁰. 

Once the pilot or project and its funding had finished, its ideas and implementation were not scalable or sustainable; they were unaffordable. Pilots and projects were usually conducted within formal educational institutions among their students. Also, evaluation and dissemination focused on technical feasibility, proof-of-concept and theoretical findings. They rarely addressed outcomes that would sway institutional managers and impact institutional performance metrics. As a result, these ideas remained optional margins of institutional activity rather than the regulated business of courses, qualifications, assessments and certificates. Nor was there a business model to support long-term adoption. 

In fairness, we should also factor in the political and economic climate at the end of the 2000s. The ‘subprime mortgage’ crisis¹¹ and the ‘bonfire of the quangos’¹² depleted the political goodwill and public finances for speculative development work. Work that had previously and implicitly assumed the ‘diffusion of innovations’¹³ into mainstream provision. That ‘trickle down’ would take these ideas from pilot project to production line.

The Shift in Mobile Learning: What Changed?

Certainly not the political or economic climate, but mobile phones are now familiar, ubiquitous and powerful, and so is artificial intelligence (AI), also familiar, ubiquitous and powerful. Both of these technologies are outside educational institutions rather than confined within them. 

These earlier pilots and projects were basically ‘dumb’ systems, with no ‘intelligence’, drawing only on information previously loaded into their closed systems. Now, we have ‘intelligence’, we have AI and we have AI chatbots on mobile phones. However, currently, AI lacks context and cannot know or respond to the location, history, activity or behaviour of the learner and their mobile phone. Unfortunately, many current AI applications and chatbots are stateless and do not retain memory across interactions, and this represents a further challenge to any continuity.

The Possibilities of Mobile Learning: Could We Do Better Now?

Today’s network technologies can enable distributed connected contributions and consumption, enabling writing and reading. These might realise more of the possibilities of constructivism and social constructivism. They could enable educational systems to learn about and respond to their individual learners and their environment, connecting groups of learners and showing them how to support each other¹⁴. 

So, is there the possibility of convergence? Is it possible to combine the ‘intelligence’ of AI, the ‘memory’ of databases and the context provided by mobile phones, including both the learner and their environment? Could this be merged and mediated by educational expertise, acting as an interface between the three technologies, filtering, selecting and safeguarding?

What might this look like? We could start by adding ‘intelligence’ and ‘memory’ to our earlier examples.

The Future of Mobile Learning: What Could it Look Like? 

In terms of formal learning, our previous examples of the Uffizi Galleries, the Lake District, the Berlin Wall and Nottingham Castle are easy to extrapolate and imagine. Subject to a mediating educational layer, learners would each be in touch with other learners, helping each other in personalised versions of the same task. They could receive background information, ideas, recommendations, feedback and suggestions, cross-referenced with deadlines, schedules and assignments from their university LMS, all based on the cumulative history of their individual and social interactions and activities. 

When it comes to community learning or visitor attractions, systems could be created that encourage interactive, informal learning. For example, a living local history or 3D community poem spread around in the air, held together by links and folksonomies¹⁵, perhaps using tags to connect ideas, a living virtual world overlaying the real one. These systems could also support more prosaic purely educational applications, combining existing literary, artistic or historical sources with personal reactions or recollections.

Technically, this is about accessing the mobile phone’s contextual data, but sometimes other simple mobile data communications, for context. It also requires querying a relational database¹⁶ to retrieve history and constraints, and perhaps an institutional LMS, to retrieve assignments, timetables and course notes. AI can then be prompted to bring these together for some educational activity. Certainly, a proof of concept is eminently feasible. The expertise and experience of the three core disciplines are still out there and only need to be connected, tasked and funded.

Conclusions and Concerns

This piece sketches some broad educational possibilities once we enlist AI to support various earlier kinds of contextual mobile learning. Specific implementations and developments must address considerable social, legal, ethical and regulatory concerns and requirements. The earlier generation of projects might have already worked with these, privacy and surveillance being the obvious ones. Still, AI adds an enormous extra dimension to these, and there are other concerns like digital over-saturation, especially of children and vulnerable adults.

Nonetheless, this convergence of AI, contextual mobile data and educational expertise promises a future where learning is not confined to traditional settings but is a fluid, intelligent and deeply embedded aspect of our daily lives, making education more effective, accessible and aligned with individual and societal needs.

1. There is considerable literature, including:
   Special editions: Research in Learning Technology, Vol. 17, 2009. 
   Review articles: Kukulska-Hulme, A., Sharples, M., Milrad, M., Arnedillo-Sanchez, I. & Vavoula, G. (2009). Innovation in mobile learning: A European perspective. International Journal of Mobile and Blended Learning, 1(1), 13–35.
   Aguayo, C., Cochrane, T. & Narayan, V. (2017). Key themes in mobile learning: Prospects for learner-generated learning through AR and VR. Australasian Journal of Educational Technology, 33(6).
   Edited books: Traxler, J. & Kukulska-Hulme, A. (Eds) (2015), Mobile Learning: The Next Generation, New York: Routledge. (Also available in Arabic, 2019.) 
   More philosophically, Traxler, J. (2011) Context in a Wider Context, Medienpädagogik, Zeitschrift für Theorie und Praxis der Medienbildung. The Special Issue entitled Mobile Learning in Widening Contexts: Concepts and Cases (Eds.) N. Pachler, B. Bachmair & J. Cook, Vol. 19, pp. 1-16. ↩︎
2. Meaning, ‘real-life’ settings. ↩︎
3. Meaning, ‘real-life’ tasks. ↩︎
4. Meaning, learning tailored to each separate individual learner.  ↩︎
5. Educational technology researchers distinguish between what teachers say, what they ‘espouse’, and what they actually do, what they ‘enact’, usually something far more conservative or traditional. ↩︎
6. An educational philosophy based on learners actively building their knowledge through experiences and interactions. ↩︎
7. A variant of constructivism that believes that learning is created through social interactions and through collaboration with others. For an excellent summary of both, see: https://www.simplypsychology.org/constructivism.html  ↩︎
8. For an explanation, see: https://smartasset.com/investing/the-economics-of-mobile-apps ↩︎
9. A common term among computing professionals, referring to whether or not different systems, such as hardware, software, applications and peripherals, will actually work together, or whether it would be more like trying to fit a UK plug into an EU socket.  ↩︎
10. A more detailed account is available at: https://medium.com/@Jisc/what-killed-the-mobile-learning-dream-8c97cf66dd3d ↩︎
11. For an explanation, see:https://en.wikipedia.org/wiki/Subprime_mortgage_crisis ↩︎
12. For an explanation, see: 2010 UK quango reforms – Wikipedia, which impacted Becta, the LSDA, Jisc and other edtech supporters.  ↩︎
13. For an explanation, see: https://en.wikipedia.org/wiki/Diffusion_of_innovations ↩︎
14. The proximity of physical or geographical context that the location awareness of neighbouring mobile phones could extend to embrace social proximity, meaning learners who are socially connected, or educational proximity, meaning learners working on similar tasks. The latter idea connects to the notions of ‘scaffolding’, ‘the more knowledgeable other’ and ‘the zone of proximal development’ of the theorist Vygotsky. For more, see: https://en.wikipedia.org/wiki/Zone_of_proximal_development ↩︎
15. Databases conventionally have a fixed structure, for example, personal details based on forename, surname, house name, street name and so on, with no choice. Folksonomies, by contrast, are defined by the user, often on the fly. For example, tagging with labels such as ‘people I like’, ‘people nearby’, ‘people with a car’. Diigo, a social bookmarking service, uses tagging to implement a folksonomy. ↩︎
16. Relational databases, unlike ‘flat’ databases based solely on a file, capture relationships, such as a teacher working in a college or a student enrolling in a course, and include all the various individual teachers, courses, students and colleges. ↩︎

About Avallain

At Avallain, we are on a mission to reshape the future of education through technology. We create customisable digital education solutions that empower educators and engage learners around the world. With a focus on accessibility and user-centred design, powered by AI and cutting-edge technology, we strive to make education engaging, effective and inclusive.

Find out more at avallain.com

About TeacherMatic

TeacherMatic, a part of the Avallain Group since 2024, is a ready-to-go AI toolkit for teachers that saves hours of lesson preparation by using scores of AI generators to create flexible lesson plans, worksheets, quizzes and more.

Find out more at teachermatic.com

_

Contact:

Daniel Seuling

VP Client Relations & Marketing

dseuling@avallain.com

The implementation of AI in education presents both opportunities and challenges. As AI solutions focused on education evolve, it’s essential to determine what constitutes ‘better’ learning. To do this, we must consider the various perspectives of teaching and learning.

Making Learning Better

Author: John Traxler, UNESCO Chair, Commonwealth of Learning Chair and Academic Director of the Avallain Lab

Understanding Different Teaching and Learning Approaches

St. Gallen, February 26, 2025 – There are many perspectives for understanding teaching and learning, each with its own values, methods and achievements. These include:

• Behaviourism focuses on observable and objective improvement in learners. Didactic and transmissive approaches that concentrate on content, absorbing information, procedures and techniques. 
• Constructivism is the belief that learning is better if it enables learners to build on their existing understandings; recognising their individuality, background, achievements and contributions.
• Social constructivism believes learning is better if learners undertake and discuss learning as a social phenomenon and group activity. It argues that learners can help each other often better than teachers, who are distant from their struggles and backgrounds.

We could characterise behaviourism as Web 1.0, where learning follows a top-down approach. In contrast, constructivism and social constructivism align more with Web 2.0, a flat, outward and collaborative approach.

There are many strategies used to deliver these perspectives, for example, quizzes, lectures, tutorials, projects, exams, workshops, role play, spaced learning, field trips, games and role-play.

The Challenge of Defining ‘Better’ in Learning

There is however always the problem of which perspective is ‘better’, and which strategy is ‘better’ for delivering it, problems without solutions. Each perspective and strategy comes with its own objectives and its own way of measuring whether those objectives are being met.  

We must however address the problem of ‘better’ since the introduction of AI into education, without considering this issue runs several risks, namely that educational AI, especially in its ‘raw’ form,

• Reinforces those perspectives based on content generation, manipulation and transmission (text, images, sound, video) because AI is good at that (as opposed to other perspectives of learning based on the learners, their individuality and interactions).
• Is justified by the ‘time-saved’ argument, de-skilling teachers or taking them out of the loop, consolidating the pedagogic status quo.
• Amplifies existing problems and inequalities beyond our capacity to deal with them.
• Struggles with the cognitive, affective and cultural diversity and individuality of learners.

The purpose of this piece is to suggest that there is another approach to the question of which perspective or strategy is ‘better’ and that is to look at it from an ethical point of view.

An Ethical Perspective on ‘Better’ Learning

Basic and widely held ethical principles talk about respect for the individual, their agency and autonomy, as well as respect for their background, culture and community, ultimately, treating them with dignity. These principles also uphold the commitment to non-maleficence and doing no harm.

If we explore different learning perspectives and strategies from this angle, then we should be asking which ones:

• Encourage curiosity, creativity, originality and criticality.
• Cause embarrassment, shame, harassment, bias or prejudice.
• Reinforce existing inequalities and divisions.
• Recognise the need to survive and flourish in a complex, changing and volatile world.
• Value humour, laughter and care, and respond to sadness or distress.
• Undermine learners’ self-confidence or self-esteem.
• Recognise their ideas and contributions.
• Treat their culture and community with respect.
• Value difference and individuality.
• Understand individual struggle and effort. 

Our systems and our technologies, perhaps mediated by teachers or perhaps supporting teachers’ good practices, should be built, evaluated, monitored and improved around these questions; these questions determine which learning is ‘better’. 

AI, Ethics and Cultural Contexts in Education

The Avallain Lab is working on these challenges, from both ends. From the bottom-up, looking at trapping and preventing individual types of harmful responses from educational AI systems, and from the top-down, looking at how educational AI systems can work with general ethical and pedagogic principles. Avallain Intelligence, our broader AI strategy, already incorporates much of this thinking in Avallain Author, Avallain Magnet and Teachermatic, shielding teachers and editors from the ‘raw’ but rather wayward and irresponsible power of AI.

There is however a complication, namely culture. Different cultures, communities, nations or societies, will have different values about:

• Individuals as opposed to the group.
• Authority as opposed to discussion. 
• Local as opposed to global.
• The future as opposed to the present or the past.
• Originality, creativity, innovation, debate and disagreement as opposed to tradition, consensus, conformity, compliance and agreement. 
• Risk-taking, chance and change as opposed to risk-avoidance, stagnation and stasis.

The Avallain Lab is focused on capturing and incorporating more of the learner’s context, including their culture and backgrounds. This approach aims to refine the responses of educational AI systems, ensuring they better align with the values and expectations of learners. At the same time, we maintain our commitment to ethical principles.

So as we continue to navigate the complexities of AI in education, it’s crucial to approach these challenges from both practical and ethical perspectives. 

About Avallain

At Avallain, we are on a mission to reshape the future of education through technology. We create customisable digital education solutions that empower educators and engage learners around the world. With a focus on accessibility and user-centred design, powered by AI and cutting-edge technology, we strive to make education engaging, effective and inclusive.

Find out more at avallain.com

About TeacherMatic

TeacherMatic, a part of the Avallain Group since 2024, is a ready-to-go AI toolkit for teachers that saves hours of lesson preparation by using scores of AI generators to create flexible lesson plans, worksheets, quizzes and more.

Find out more at teachermatic.com

_ 

Contact:

Daniel Seuling

VP Client Relations & Marketing

dseuling@avallain.com 

The UK’s Department for Education publishes its outcomes-oriented safety recommendations for GenAI products, addressed to edtech companies, schools and colleges.

UK’s Generative AI: Product Safety Expectations

Author: Carles Vidal, Business Director of the Avallain Lab

St. Gallen, February 20, 2025 – On 22 January 2025, the UK’s Department for Education (DfE) published its Generative AI: Product Safety Expectations. This is part of the broader strategy to establish the country as a global leader in AI, as outlined in the Government’s AI Opportunities Action Plan. 

As a leading edtech company with over 20 years of experience, Avallain was invited to participate in consultations on the Safety Expectations. Avallain Intelligence’s focus on clear ethical guidelines for safe AI development, demonstrated through TeacherMatic and other AI-driven solutions across our product portfolio, is reflected in our role in these consultations, where we were well-positioned to contribute expert advice.

Product Expectations for the EdTech Industry

The Generative AI: Product Safety Expectations define the ‘capabilities and features that GenAI products and systems should meet to be considered safe for use in educational settings.’ The guidelines, aimed primarily at edtech developers, suppliers, schools and colleges, come at a crucial time. Educational institutions need clear frameworks to assess the trustworthiness of the AI tools they are adopting. The independent report, commissioned by Avallain, Teaching with GenAI: Insights on Productivity, Creativity, Quality and Safety, provides valuable insights to help inform these decisions and guide best practices.

Legal Alignment, Accountability and Practical Implementation

The guidelines are specifically intended for edtech companies operating in England. While not legally binding, the text links the product expectations to existing UK laws and policies, such as the UK GDPR, Online Safety Act and Keeping Children Safe in Education, among others. This alignment helps suppliers, developers and educators navigate the complex legal landscape. 

From an accountability point of view, the DfE states that, ‘some expectations will need to be met further up the supply chain, but responsibility for assuring this will lie with the systems and tools working directly with schools and colleges.’ Furthermore, the guidelines emphasise that the expectations are focused on outcomes, rather than prescribing specific approaches or solutions that companies should implement.

Comparing Frameworks and An Overview of Key Categories

In line with other frameworks for safe AI, such as the EU’s Ethics Guidelines for Trustworthy AI, the Generative AI: Product Safety Expectations are designed to be applied by developers and considered by educators. However, unlike the EU’s guidelines, which are field-agnostic and principles-based, the DfE’s text is education-centred and structured around precise safety outcomes. This makes it more concrete and focused, though it is less holistic than the EU framework, leaving critical areas such as societal and environmental well-being out of its scope.

The guidance includes a comprehensive list of expectations organised under seven categories, summarised in the table below. The first two categories — Filtering and Monitoring and Reporting — are specifically relevant to child-facing products and stand out as the most distinctive of the document, as they tackle particular risk situations that are not yet widely covered.

The remaining categories — Security, Privacy and Data Protection, Intellectual Property, Design and Testing and Governance — apply to both child- and teacher-facing products. They are equally critical, as they address these more common concerns while considering the specific educational context in which they are implemented.

Collaboration and Future Implications

By setting clear safety expectations for GenAI products in educational settings, the DfE provides valuable guidance to help edtech companies and educational institutions collaborate more effectively during this period of change. As safe GenAI measures become market standards, it is important to point out that the educational community also needs frameworks that explore how this technology can foster meaningful content and practices across a diverse range of educational contexts.

Generative AI: Product Safety Expectations — Summary

• Filtering
  1. Users are effectively and reliably prevented from generating or accessing harmful and inappropriate content.
  2. Filtering standards are maintained effectively throughout the duration of a conversation or interaction with a user.
  3. Filtering will be adjusted based on different levels of risk, age, appropriateness and the user’s needs (e.g., users with SEND).
  4. Multimodal content is effectively moderated, including detecting and filtering prohibited content across multiple languages, images, common misspellings and abbreviations.
  5. Full content moderation capabilities are maintained regardless of the device used, including BYOD and smartphones when accessing products via an educational institutional account.
  6. Content is moderated based on an appropriate contextual understanding of the conversation, ensuring that generated content is sensitive to the context.
  7. Filtering should be updated in response to new or emerging types of harmful content.
  8. Filtering should be updated in response to new or emerging types of harmful content.
• Monitoring and Reporting
  1. Identify and alert local supervisors to harmful or inappropriate content being searched for or accessed.
  2. Alert and signpost the user to appropriate guidance and support resources when access to prohibited content is attempted (or succeeds).
  3. Generate a real-time user notification in age-appropriate language when harmful or inappropriate content has been blocked, explaining why this has happened.
  4. Identify and alert local supervisors of potential safeguarding disclosures made by users.
  5. Generate reports and trends on access and attempted access of prohibited content, in a format that non-expert staff can understand and which does not add too much burden on local supervisors.
• Security
  1. Offer robust protection against ‘jailbreaking’ by users trying to access prohibited material.
  2. Offer robust measures to prevent unauthorised modifications to the product that could reprogram the product’s functionalities.
  3. Allow administrators to set different permission levels for different users.
  4. Ensure regular bug fixes and updates are promptly implemented.
  5. Sufficiently test new versions or models of the product to ensure safety compliance before release.
  6. Have robust password protection or authentication methods.
  7. Be compatible with the Cyber Security Standards for Schools and Colleges.
• Privacy and Data Protection
  1. Provide a clear and comprehensive privacy notice, presented at regular intervals in age-appropriate formats and language with information on:
  2. The type of data: why and how this is collected, processed, stored and shared by the generative AI system.
  3. Where data will be processed, and whether there are appropriate safeguards in place if this is outside the UK or EU.
  4. The relevant legislative framework that authorises the collection and use of data.
  5. Conduct a Data Protection Impact Assessment (DPIA) during the generative AI tool’s development and throughout its life cycle.
  6. Allow all parties to fulfil their data controller and processor responsibilities proportionate to the volume, variety and usage of the data they process and without overburdening others.
  7. Comply with all relevant data protection legislation and ICO codes and standards, including the ICO’s age-appropriate design code if they process personal data.
  8. Not collect, store, share, or use personal data for any commercial purposes, including further model training and fine-tuning, without confirmation of appropriate lawful basis.
• Intellectual Property
  1. Unless there is permission from the copyright owner, inputs and outputs should not be:
     • Collected
     • Stored
     • Shared for any commercial purposes, including (but not limited to) further model training (including fine-tuning), product improvement and product development.
  2. In the case of children under the age of 18, it is best practice to obtain permission from the parent or guardian. In the case of teachers, this is likely to be their employer—assuming they created the work in the course of their employment.
• Design and Testing
  1. Sufficient testing with a diverse and realistic range of potential users and use cases is completed.
  2. Sufficient testing of new versions or models of the product to ensure safety compliance before release is completed.
  3. The product should consistently perform as intended.
• Governance
  1. A clear risk assessment will be conducted for the product to assure safety for educational use.
  2. A formal complaints mechanism will be in place, addressing how safety issues with the software can be escalated and resolved in a timely fashion.
  3. Policies and processes governing AI safety decisions are made available.

About Avallain

At Avallain, we are on a mission to reshape the future of education through technology. We create customisable digital education solutions that empower educators and engage learners around the world. With a focus on accessibility and user-centred design, powered by AI and cutting-edge technology, we strive to make education engaging, effective and inclusive.

Find out more at avallain.com

About TeacherMatic

TeacherMatic, a part of the Avallain Group since 2024, is a ready-to-go AI toolkit for teachers that saves hours of lesson preparation by using scores of AI generators to create flexible lesson plans, worksheets, quizzes and more.

Find out more at teachermatic.com

_ 

Contact:

Daniel Seuling

VP Client Relations & Marketing

dseuling@avallain.com