Digital Tech and Mountain Out-Migration

17 March 2026 ~ Richard Heeks ~ Leave a comment

What role does digital technology play in the depopulation of mountain regions?

This question matters because mountain communities worldwide face accelerating out-migration, particularly of young people. While this “can help to reduce poverty and diversify livelihoods in mountains and beyond” it “has reached such a scale that depopulation and the seasonal absence of people of working age are widespread” with potential negative impacts “for the lives of those who stay behind, for the social fabric of mountain communities, and for the management of mountain ecosystems”[1].

From this source and others, the dynamics of mountain out-migration are well-established. Limited economic opportunities, inadequate services and infrastructure, and the physical challenges of mountain life – increasingly exacerbated by the impact of climate change – drive people to leave. But digital tech adds some complexities to this established pattern, from which I’ll explore three potential roles it can play: accelerator, enabler, and counter-force.

Digital accelerating out-migration

That problem of “inadequate services and infrastructure” applies to digital, with digital divides of both connectivity and skills being particularly strong for remote mountain communities. These divides limit access to education, enterprise and public services, and make urban migration more attractive.[2] For mountain youth, exposure through social media to urban lifestyles and opportunities can amplify dissatisfaction with limited local prospects, as they get an immediate – if distorted – sense of what life elsewhere looks like.[3]

Digital supporting out-migration

Digital connectivity supports out-migration by making it more sustainable for individuals and communities. Mobile money and digital remittance platforms reduce transaction costs for migrants supporting families back home[4]. Video and audio calling maintains social bonds across distance in ways that were impossible a generation ago[5]. This digital infrastructure of migration means that leaving no longer requires the same degree of social rupture. For mountain communities, this might preserve some economic and social links, but it also removes friction that might otherwise encourage people to stay or return.

Digital reversing out-migration

There is evidence that improved digital connectivity such as broadband can increase the reach, growth and employment of enterprises in rural areas[6]. Whether this could be linked to a reduction in out-migration is, however, questionable because of the strength of the accelerator and enabler effects. One study from Spain, for example, finds “no evidence of a causal effect” between broadband growth and rural population change[7], while a similar study in China found “broadband creates ‘digital routes’ facilitating outmigration rather than ‘digital roots’ anchoring residents to rural areas”[8].

There’s a similar picture with the growth of interest in remote working; the idea that local people could undertake digital work in mountain communities, or even attract in-migration of digital nomads[9]. While this is both feasible and happening in some global North mountain communities, and digital nomads are setting up in well-connected beach resorts in the global South, there remain serious barriers to this as a strategy for most global South mountain regions[10].

Policy Priorities

A first priority is to strategically position mountain connectivity as integrated within a broader development architecture. This means making services affordable and resilient, e.g. during extreme weather conditions. It also means developing the skills to make productive use of the technology. And it means recognising that connectivity alone does not reverse migration without complementary economic development and service provision.

A second priority is to envisage migration patterns as circular. This means designing digital tools to help maintain productive links between mountain communities and those who have migrated out, potentially facilitating return migration or continued economic contribution such investment in local productive infrastructure and enterprise.

Of course, all this assumes that reversing mountain depopulation is a worthwhile goal and that abandonment to nature of (more remote) mountain regions should not be the intent.

Research Priorities

For researchers, key gaps remain around the actual versus assumed impact of digital connectivity on migration decisions in mountain contexts. Much current understanding is extrapolated from lowland rural areas. We lack longitudinal, causal evidence on which forms of digital investment genuinely reduce out-migration, which ones mainly facilitate mobility outwards, and which do both. Understanding these dynamics requires mixed methods approaches that can capture both structural factors and individual decision-making across different mountain regions and cultural contexts.

A second gap is more political: research has not kept pace with the governance questions surrounding digital roll-out in mountain regions, including who bears the costs and who captures the gains since will also shape patterns of migration. We thus need to know – alongside economic, social, and environmental impacts – what are the political implications of digital’s increasing interaction with mountain land, environment and labour markets.

Originally published in the Mountain Digital Futures newsletter: https://www.linkedin.com/newsletters/7400288732999733248/

Image source: https://migrationrightslab.org/migration-and-gender-in-west-africa-ghana-brief/

[1] https://adaptationataltitude.org/wp-content/uploads/2023/05/bachmann_et_al_migration_and_smd_2019_low_0.pdf

[2] https://www.defindia.org/wp-content/uploads/2023/07/Connecting-Himalayan-Communities-An-Issue-Brief_PRINT.pdf. For rural areas more generally, see: https://academic.oup.com/jcmc/article/21/3/247/4065369 and https://www.oecd.org/en/data/insights/statistical-releases/2025/07/digital-connectivity-expands-across-the-oecd-but-rural-areas-are-falling-further-behind.html

[3] https://www.researchgate.net/publication/399601704_Rural-Urban_Migration_of_Young_People_in_High_Andean_Communities_in_Peru_Imaginaries_and_Practices_of_Vulnerability_and_Social_Advancement. Noting that social media is just continuing a longer historical trend of media-based exposure to urban lifestyles: https://onlinelibrary.wiley.com/doi/10.1111/j.1468-2435.2010.00627.x

[4] E.g. https://migrantmoney.uncdf.org/resources/insights/integrating-remittance-and-mobile-wallet-services-a-case-study-of-ime-pay-in-nepal

[5] https://ict4d.org.uk/wp-content/uploads/2023/09/ict4d-rwp-1-nepal-v5-1.pdf

[6] https://ruralinnovation.us/press-releases/new-research-proves-that-providing-fiber-broadband-experiences-to-rural-communities-boosts-income-entrepreneurship-and-business-investment/; https://www.sciencedirect.com/science/article/abs/pii/S0743016725003237; https://publications.iadb.org/en/access-credit-and-expansion-broadband-internet-peru

[7] https://onlinelibrary.wiley.com/doi/10.1111/tesg.12596

[8] https://www.iza.org/publications/dp/17752/digital-roots-or-digital-routes-broadband-expansion-and-the-rural-urban-migration-in-china

[9] https://pub.nordregio.org/r-2024-7-remote-work-in-rural-areas/r-2024-7-remote-work-in-rural-areas.pdf

[10] https://pmc.ncbi.nlm.nih.gov/articles/PMC11387329/; https://wol.iza.org/articles/does-working-from-home-work-in-developing-countries/long

From Willingness to Governance: Why Cyber Threat Intelligence Sharing Becomes Circumstantial

4 March 2026 ~ M.A Nainna ~ 1 Comment

This commentary reflects on the paper “Factors Amplifying or Inhibiting Cyber Threat Intelligence (CTI) Sharing,”, which reports a qualitative, grounded-theory-informed study based on nine semi-structured interviews with UK cybersecurity professionals. The paper focuses on the external sharing of CTI actionable information such as indicators of compromise (IOCs), tactics/techniques, and incident context between organisations (e.g., peers, sector groups, vendors, and public agencies). It examines how practitioners interpret the risks, incentives, and organisational constraints surrounding CTI sharing, and it offers a behavioural framing of sharing—enthusiastic, circumstantial, and unenthusiastic—which helps explain why sharing often becomes conditional rather than routine.

Read the paper: https://doi.org/10.1007/978-3-031-56481-9_14

The most useful move is not simply re-listing “barriers and enablers” (which the literature already does) but classifying sharing behaviour into three empirically derived patterns: enthusiastic, circumstantial, and unenthusiastic CTI sharing.

The enthusiastic group is described through mechanisms that look institutional and procedural: government-led dissemination channels, defined processes (e.g., policing/national intelligence practices), and practical awareness of regulatory responsibilities such as the General Data Protection Regulation (GDPR). These are not “warm feelings” about collaboration; they are routinised structures that make sharing normal, legible, and defensible.

The headline theoretical contribution is circumstantial sharing. In this mode, practitioners may perform rigorous investigation and mitigation work and may even recommend external disclosure; however, disclosure becomes contingent on organisational dynamics, especially senior management decision-making. This reframes CTI sharing away from an individual-level “willingness” narrative and toward a governance narrative: “Who has the mandate to share, and under what conditions?”

By contrast, the unenthusiastic group is anchored in political economy. Participants link reluctance to the capitalist nature of the cybersecurity market, weak inter-organisational trust, and the perceived risk that shared intelligence will advantage competitors or even adversaries. The paper is strongest when it treats this as rational behaviour under market incentives, rather than a moral failure to collaborate.

Figure 1. Conceptual map of CTI sharing behaviours factors and key impediments

The “impediments” also land with practical clarity: fear of regulatory penalties, cost (both paying for CTI and operationalising sharing), and competitive/adversarial exposure. Notably, the language/standardisation issue (“a way of describing things that everyone agrees on”) adds an interoperability layer: even when organisations want to share, they may not share effectively.

Summary table: behaviours factors and drivers (from the paper)

CTI sharing behaviour	What it looks like in practice	Main amplifiers/inhibitors highlighted
Enthusiastic sharing	External sharing proceeds readily via established channels	Government initiatives; defined processes, and regulatory awareness (e.g., GDPR)
Circumstantial sharing	Practitioners investigate and escalate; sharing depends on management and context	Fear of regulatory penalties, cost, lack of shared language, competitor/adversary risk, strict controls (e.g., policing)
Unenthusiastic sharing	External sharing avoided or heavily resisted	Low trust; commercial/market dynamics; reluctance to aid competitors/adversaries

A constructive critique is that the study’s empirical base (nine UK interviews) is well suited to theory-building, but the paper could more explicitly separate (a) inhibitors of decision-to-share, (b) inhibitors of ability-to-share, and (c) inhibitors of value-from-sharing. That would sharpen how frameworks like Structured Threat Information eXpression (STIX), Trusted Automated eXchange of Intelligence Information (TAXII), Malware Information Sharing Platform (MISP) and governance (policies, liability, reporting lines) interact, rather than sitting in parallel sections.

One way to operationalise this critique is to reorganise the reported impediments into three analytically distinct stages of inhibitors in the sharing pipeline: (a) what prevents authorisation to disclose (decision-to-share), (b) what prevents effective packaging and transmission (ability-to-share), and (c) what prevents recipients from realising actionable benefit (value-from-sharing), which also makes it easier to see where governance dominates versus where technical frameworks add leverage.

(a) Inhibitors of decision-to-share

These factors block the “authorisation” step: even when analysts want to share, the organisation chooses not to.

Liability and penalty risk: fear of regulatory consequences (and uncertainty about what is permissible) pushes leaders toward risk avoidance.

Reputational exposure: concern that disclosure signals weakness, triggers customer agitation, or invites scrutiny.

Competitive/adversary calculus: worry that sharing benefits competitors or provides adversaries with insight into detection/response posture.

Governance bottlenecks: unclear reporting lines, slow approvals, or management override (“management decides whether to disclose”).

Framework–governance interaction: standards like STIX/TAXII don’t resolve “who” can approve sharing or “what” legal basis is acceptable; policies, liability models, and decision rights dominate here.

(b) Inhibitors of ability-to-share

These factors block the “execution” step: the organisation may be willing but cannot share efficiently or safely.

Lack of shared language / weak standardisation in practice: inconsistent vocabularies, formats, and levels of abstraction make “good CTI” hard to package.

Process immaturity: absence of repeatable workflows for sanitisation, classification, Traffic Light Protocol (TLP), de-identification, and release.

Tooling and integration gaps: limited capability to structure, transport, and operationalise CTI (where STIX/TAXII/MISP would normally help).

Resource constraints: time, staffing, and financial costs of producing usable intelligence for others.

Framework–governance interaction: STIX/TAXII/MISP increase “technical interoperability”, but they depend on governance primitives (data classification rules, sanitisation policy, ownership, accountable roles) to function reliably.

These factors undermine the “payoff” step: even if CTI is shared, recipients may not get actionable benefit, reducing motivation to sustain sharing.

Low signal-to-noise / questionable relevance: shared CTI may be too generic, too late, or not aligned with the recipient’s context.

Poor actionability: indicators without context (TTPs, confidence, provenance, mitigation guidance) don’t translate into defensive action.

Asymmetric exchange: perceptions of “free riding” or unequal benefit erode trust and long-term participation.

Commercial data control: when CTI is treated as a product, access limitations and licensing constraints reduce collective value creation.

Framework–governance interaction: structured formats can improve enrichment (confidence, provenance, relationships), but “value” ultimately depends on community norms, reciprocity rules, and outcome feedback loops (what helped, what didn’t).

In short, governance determines “whether” sharing happens (decision-to-share), frameworks and operational controls determine “how” it happens (ability-to-share), and community norms plus CTI quality determine “why” it is sustained (value-from-sharing).

Call to action: If you work on digital development, policy, or organisational governance, consider auditing your CTI-sharing pipeline against the three stages (decision-to-share, ability-to-share, value-from-sharing) and identify where governance, process, or tooling is the binding inhibitor. You can read the paper via the DOI link above, and I’m happy to discuss practical implications or collaboration opportunities. Feel free to email me a.n.muhammad@edu.salford.ac.uk or message me on LinkedIn Abubakar Muhammad Nainna, PhD

Using Design–Reality Gap Analysis to Strengthen Computing/ICT Curriculum Development

24 February 2026 ~ Richard Heeks ~ Leave a comment

Given rapid and continuous change in digital tech, universities face pressure to keep their curricula relevant. But updating computing programmes – computer science, information systems, ICT, etc – is rarely straightforward.

Over the years, I’ve seen many institutions commit to curriculum reform only to find that intended changes don’t materialise in practice, or that ambitious designs encounter stubborn institutional obstacles.

Julian Bass and I explored this problem through a detailed study of curriculum reform in Ethiopian universities, published open access in the Electronic Journal of Information Systems in Developing Countries. While the case itself offers useful insights, the value of the paper lies in something more general: demonstrating how design–reality gap analysis provides a systematic way to assess and manage curriculum change. The model is relevant for anyone leading undergraduate or postgraduate ICT-related programmes.

Why a design–reality gap approach?

Most curriculum reforms begin with a well-intentioned design: new competencies, updated course structures, industry-aligned learning outcomes, more practical assignments, and so on. Yet these designs rest on a set of assumptions about available staff expertise, technology, processes, resources, and institutional culture. When these assumptions collide with reality, gaps emerge. And if these gaps are large enough, they undermine implementation no matter how strong the curriculum looks on paper.

The design–reality gap model helps educators and managers identify and measure these mismatches. It offers an eight-part framework, summarised under the acronym OPTIMISM that covers: objectives, processes, technology, information, management systems, investment resources, staffing/skills, and the wider milieu. In practice, this creates a structured diagnostic lens that makes implicit problems explicit.

What does this enable in practice?

The model proved extremely useful in our study but its real utility extends far beyond a single country or context. For curriculum designers and programme leaders, the approach supports at least four practical uses:

Risk identification: Before launching a new curriculum, the model highlights where the biggest vulnerabilities are likely to lie, be they in insufficient labs, limited staff experience in new subjects, misaligned assessment processes, misaligned managerial values, etc.
Prioritisation: Not all gaps matter equally. The framework helps institutions decide where to invest first, and which changes are prerequisites for others.
Realistic planning: By assessing gaps, universities can build phased implementation strategies that match existing capacity rather than relying on overly optimistic assumptions.
Iterative improvement: Curriculum reform is never a one-off event, and the model supports periodic review, enabling departments to track progress and recalibrate as conditions evolve.

Why this matters

Universities face resource constraints, rapid technological change, and growing expectations from students and employers. At the same time, new pressures particularly around AI but also in relation to cybersecurity, data ethics, digital sustainability and a slew of emergent issues, are reshaping the skills landscape. Under these conditions, a structured tool for analysing implementation challenges is more essential than ever.

Curriculum reform succeeds not simply because we articulate a compelling vision of what students should learn, but because we understand the institutional realities within which that vision must be delivered. The design–reality gap model won’t solve those challenges on its own but it offers a clear, practical framework for navigating them.

For colleagues planning updates to computing or ICT programmes, I hope this model provides a useful starting point. It’s a reminder that successful curriculum change begins not just with design but also with an honest assessment of the reality into which that design is being introduced.

Is something big happening in Research?

18 February 202618 February 2026 ~ matthaikin ~ Leave a comment

Matt Shumer’s “Something Big Is Happening” has gone viral – and understandably so. His core observation is hard to argue with: AI capability – most notably in coding but also many other fields – has crossed a threshold. The acceleration is observable and, for many practitioners, already reshaping daily work so they are literally guiding rather than doing what their jobs were 6 months ago.

He doesn’t say much about research. There might be a reason for that. Research lacks objective outputs to test against; evidence is inherently contested, differing interpretations are unavoidable, judgement calls are entangled with politics and power.

But something is definitely happening with AI-assisted research – I want to describe some of my own recent experiments with AI-assisted practice, where my role as researcher is changing, and why I’m simultaneously excited and worried about what it may mean for my work.

An illustrative walkthrough – AI-assisted rapid literature review

Landscape analyses for strategy research in international development are common. It is worth being clear that I am not (yet anyway) discussing robust and rigorous academic literature reviews, but the sort of snapshots that kick-off most UN/NGO type research to ground everyone in the current state of play for the topic being researched.

I have seen the value and role of AI tools literally transform in the last 3-6 months across every stage:

Sourcing: The first time I tried using AI to generate sources for a rapid lit review – it helped – a little… But the results were patchy – at least 20% of its references didn’t even exist and it needed significant time to review, validate and plug the gaps in the rest.

Doing the same experiment last week – entirely invented references were close to zero. The results are materially stable and usable. It still needed validation and review of course – whittling its ~80 or so suggestions down to 30-40 that are relevant, aren’t vendor promotion, or 1-page promotional work, or inaccessible paywalled sources, or just entirely tangential. What counts as relevant evidence remains a hugely important and human analytical judgement.

Sourcing that took me 3-4 days before, is now down to roughly half a day.

Synthesis & drafting: A few months back, this was still a 90% manual process – sure I’d use AI tools to explore questions, to critique drafts, to write first drafts of specific extracts – but if I tried to get it to do more, the results were so poor it probably took more time to review and correct than to just do it myself.

Now, I can get a pretty decent first (internal use only) version of a large corpus in a day (maybe a day more if I need to familiarise myself with the subject matter first, which AI also helps with), and a good enough client-facing draft in another day or two. This was more like 2-3 weeks in the past – even for the ‘non-academic’ level of rigour I am working within. Why the difference?

The latest frontier models really are just way better – I won’t lie, probably half the improvements are that simple (results from Perplexity now it has Claude Opus 4.6 under the hood are incomparable to it results before), but the rest is a more reflective way of using these models.

Research needs complex hybrid workflows
I took the time to design a careful step-by-step workflow with researcher intervention at numerous points to guide, to flag gaps, add my insights, to inject domain-specific interpretation, to warn off dead-ends etc. – what I’ve been referring to as HGTL instead of HITL (Human Guiding the Loop, not Human In the Loop which I always felt was downplaying the role of the human!). The difference is not just semantics – the human is structuring epistemic boundaries, not just approving outputs. Guiding the loop is a meaningful difference to how things seem to be developing in other types of work. Domain and methodology knowledge remain crucial and are what holds the process together. Without them, flawed studies are accepted authoritative, personal blogs risk being treated as robust evidence, poorly designed or vendor claims masquerade as research. To be fair, junior researchers often make similar mistakes – my role now feels closer to a traditional oversight role than the ‘OMG what is this mess I need to clean up’ of the not so distant past!
Model triangulation and comparison
Crucially, I learned to build Triangulation into my workflow. This is a core safeguard and probably the main reason that early AI-produced syntheses are becoming usable. Basically, I run the same prompts on the same corpus through two different frontier models and then get AI to compare the outputs before it does anything with them. Each model fails differently. One produces narrative completeness but irons out disagreement. Another surfaces tensions well but has significant gaps in what it covers. This structured comparison stage seems to force the divergence to be taken into account when the model is later asked to synthesise the two, giving markedly better initial results.

I find AI now behaves less like a fabricator or an over-eager helper, and more like an enthusiastic junior researcher. It still surfaces some irrelevant, weak, or tangential content; it missed some issues around methodological robustness; mistakes still filter through into drafts – but my level of filtering and review has shifted from fiction and damage control to improving relevance and quality. That is huge. It does not eliminate the need for someone who knows the terrain, but it changes the nature of that person’s role in meaningful ways.

What does this mean for research as a role?

For me personally, task-level gains are real and substantive, reducing the time I need to get to key milestones, and making the type of agile research I’ve always tried to do, a reality – why not do some analysis on 3-4 areas and then decide which to focus on, if each can be done in a few days – before this was a nice idea but each area would need weeks, so it rarely happened.

But there is strong emerging evidence suggesting task-level efficiency does not necessarily translate to job-level or system-level productivity and sometimes quite the opposite – review burdens shift to senior staff, expectations increase, workload becomes increasingly mismatched with pay. It is definitely not fair to say “AI makes researchers jobs faster or better”.

There is also the meta- level worry about what this new way of working may mean from a neuro-science perspective. There is a growing body of research on the idea that AI can lead to cognitive burnout. Might the very aspects I am increasingly delegating to AI (the sourcing, the initial direct engagement with the material, the pattern identification etc) also be the pieces that are keeping my brain working. What might the reduction in the slow work of reading widely, noticing what doesn’t fit, scribbling notes in the margins etc mean gets lost – for example, might the researcher’s instincts for weak evidence start to stagnate if this function is not exercised.

Or maybe the opposite is true, maybe by operating at a higher level, our brains will become better at spotting patterns, maybe having genuine back-and-forth exchanges with an AI about the material in a corpus will prove to be better for our cognition than reading it alone.

I don’t know. And research is such a different space to the areas being studied (coding, finance, admin, law etc) that it may be a long time until anyone does know.

Power dynamics – a thorn in AI’s side or can it help us see them?

Most of my work involves thinking about power structures, dynamics, locally-led approaches, Southern voices etc.

How well did this approach cope with these additional complications?

It is well documented that AI tends to reproduce structural power biases – training data reflects existing hierarchies of voice, funding, and institutional authority; global South research, practitioner knowledge, community perspectives are systematically under-represented.

I’ve found that – with the latest models – explicitly building this lens into every prompt helps more than I expected and more than I’ve seen before. Obviously it doesn’t solve it – the core issue is the training data – but it definitely mitigates it to an extent.

At the same time, I found deliberately asking the AI tools about power dynamics in a literature base, can actually help me surface things I might not have noticed – which institutional voices are over-represented, which geographies or methodological traditions are missing. It is not perfect, but the trade-offs are now real – where 6 months ago I would simply assume that all AI-produced interim outputs would be missing this power lens, now I can see it more as a thought partner in identifying them (tempered always with the knowledge that its own data and algorithms are also a major part of the problem!)

And of course, the truly interesting part of power dynamics – not just surfacing them, but the interpretive step – reframing, choosing what to foreground, what to recommend etc – stays entirely with me as the human researcher.

Is Shumer right? Is AI coming for research too?

Something big is definitely happening in research. As Shumer notes for coding and other areas – recent acceleration is real and observable and in the next six months it is likely the acceleration will be even faster

But it does seem slower and trailing behind areas like coding – likely due to subjective outputs and lack of tight feedback loops much of his argument depends on.

The trajectory I see is toward increasing delegation of stages, not toward fully autonomous research. Reconfiguration, not replacement. For now anyway.

What do I think of this direction of travel? I’m not yet sure. This article is an observation of what is already happening, not a comment on whether that is good or bad!

For me personally, efficiency gains are real and transformative at the task level. What this means at the job role or system level is yet to be seen.

If it turns out that the stages being delegated are the very stages that build the judgement to do the work well, then the question is not whether AI can do more of the research but whether we still have the expertise to know when it has done it badly.

(Note: This was first published on my LinkedIn newsletter ‘The digital power shift’ on Feb 16 2026)

A couple of links for context:

1. Harvard piece on potential negative impact of AI on job efficiency and workload: https://hbr.org/2026/02/ai-doesnt-reduce-work-it-intensifies-it

2. Summary of state of research on links between AI usage and cognitive decline: https://www.ie.edu/center-for-health-and-well-being/blog/ais-cognitive-implications-the-decline-of-our-thinking-skills/

Reading the digital economy for “difference”: From “limits” to “possibilities”

3 February 2026 ~ Anuradha Ganapathy ~ Leave a comment

Why are we more adept at de-constructing, de-bunking and de-stabilising, than re-building, re-constructing, re-booting?

This is a question I explored in a blogpost last year, asking if we could move beyond problematisation and use “critique” as a tool to reconstruct alternate hopeful possibilities (Ganapathy, 2025). I believe that the question of “how” to do this is also a project of cultivating the capacity to “read for difference”, a concept that feminist geographers JK Gibson Graham developed as a counter to the more common practice of “reading for dominance” (Gibson-Graham, 2006).

When we read for dominance, we are more likely to activate a “shut down” mode.

Gibson-Graham (2006) coined the team “reading for dominance” as a way of explaining the obduracy of capitalist dynamics in theorising economic relations. They argued that most economies inhabit multiple forms of labour practices – housework, factory work, volunteer work, as well as multiple types of employment – self-employed, family help, reciprocal labour, etc., yet we tend to position these with respect to capitalism, i.e., as being adjacent or opposite to capitalism, or as existing in capitalism’s orbit as its substandard imitation (Gibson-Graham, 2008). In other words, “reading for (capitalist) dominance” is a particular form of theorising that gives meaning to economic activities and identities only with reference to the master economic signifier of capitalism (Gibson-Graham, 2006). It functions as a form of “strong” theory by lodging its faith in a particular theorisation of the economy as a stable and self-reproducing structure, rather than seeing it as a zone of cohabitation of multiple economic forms (Ibid).

In theorisations of our digital world today, we see a strong tendency to read for dominance, e.g., in articulations of dominance of Big-tech (Hendrikse et al., 2022), or Alt-Big-tech (Parsheera, 2022), producing what has been called “Big critique” (Burgess, 2023) i.e., scholarly work that is set up principally against the unprecedented power that technology affords to technology companies or the state.

Reading for dominance stems from and feeds into a strong theory of power – i.e., power as a totalising hegemonic force that is hard to sidestep or upturn (Gibson-Graham, 2006). In this view, “everyday” or “alternative” forms of politics do little to shift its otherwise intractable presence. For example, a strong theory of platform capitalism is less likely to accommodate alternatives that do not fundamentally challenge the monopolistic structures of the current digital market regimes (Rikap, 2024), and therefore runs the risk of de-legitimising the many modest constraint-driven innovations across the Global South (Kapur, 2025) as being already co-opted, or as non-credible alternatives to the current dominant model. This then becomes a self-reinforcing theory of power (Gibson-Graham, 2006). In many ways, reading for dominance pre-disposes us to “shut-down” rather than “open-up” possibilities, because in such a reading of power, there is little that is untouched by its organising logic (Ibid).

Opening up possibilities by reading for “difference”: Two examples

Gibson-Graham (2006) argue that if we want to imagine other worlds, we must learn to read for difference, i.e., open up avenues to think beyond a singular monolithic understanding of power. In reading for difference, we challenge ingrained alignments of power that foreclose the potential for multiple trajectories of possibility to materialise (Ibid). Rather than view power as a totalising, all-or-none force, we can “de-exoticize” it, seeing it as pervasive and uneven, as something that can be enforced, co-opted, manipulated, appropriated, and resisted, giving rise to multiple zones of contested possibilities, and importantly, multiple legitimate ways of being in the world (Gibson-Graham, 2008). I offer two illustrative examples of how reading for difference v/s dominance can be materialised in the context of the digital economy. These are not comprehensive or definitive applications of the concept; they only represent instantiations of how we can think with it.

Example 1:

In 2015, Facebook launched a service called Free Basics in India, which provided mobile users with access to a small number of websites, free of data charge. It was fiercely contested as a form of digital colonialism, a “walled garden” that furthered Facebook’s own business objectives through the collection of user data, rather than providing true internet access (Pahwa, 2015). Eventually, following pressure from digital rights activists and civil society organisations, it was banned in India (BBC, 2006). However, Free Basics’ defeat in India did not stop the project’s expansion across the African continent, where it was defended by digital rights activists and civil society organisations (Nothias, 2020). Against the backdrop of increasing government control over the Internet, they argued that Free Basics could be a useful gateway to the Internet for first time and price sensitive users and saw it as part of a longer tradition of free packages offered by telecom service providers (Ibid).

When we read for dominance, we are more likely to discredit Free Basics as a quintessentially capitalist project that fuels corporate ambition. This argument, while not wrong, does not make room for “bargains with the devil” and is therefore also less likely to be able to account for real-world transgressions and the murkiness of everyday politics (Gibson-Graham, 2006). When we read for difference though, we are more likely to develop the explanatory potential to describe how interactions between states, transnational corporations, civil society organisations and citizens may produce vastly different responses to “capitalist projects”, or indeed how capitalist alternatives may be perceived as more favourable in certain contexts which also represents how power materialises in the real world.

Example 2:

Namma Yatri is a mobility solution in Bengaluru, India that came out of a collaboration between multiple entities: a payments startup, a nonprofit, and a local drivers’ union. While it is a privately held for-profit company, workers unions were involved in the design and implementation of its mobility solution (Namma Yatri, n.d.). Built on open-source protocols, Namma Yatri follows a zero-commission model, charging low-cost subscription fee for drivers; this enables higher earnings and more control over the amounts (Chaudhary, 2025). Furthermore, Namma Yatri also publishes driver data on rides as well as earnings in order to promote transparency of operations (Namma Yatri, n.d.). It is hard to fit Namma Yatri neatly into either end of the profit-centric (exploitative, extractive) or worker-centric (worker-owned, state backed) spectrum.

A reading for dominance will likely see this as an extension of the former, arguing that it ultimately does little to reverse power dynamics between the those who control and run the platform infrastructure v/s those who do not. A reading for difference, on the other hand will likely see it as a model that has altered the market dynamics in a way that is potentially more favourable to workers. For example, Namma Yatri has quietly disrupted the commission-based structures of leading platforms such as Uber and Ola, pushing them to move to zero-commission models (Aravind, 2025). However, the core question of whether workers themselves see this as a fairer alternative still remains hard to answer because workers actively and strategically multi-home, refusing affiliation or loyalty to any one platform (Chaudhary, 2025). Any connections, therefore, between alternate platform models, market behaviours and worker experiences can only be made cautiously by attending to the specificity and multiplicity of these arrangements and experiences, which we are better placed to do when we read for difference.

Of course, none of the conclusions drawn here are novel or revolutionary by themselves. The point here is to show that reading Free Basics or Namma Yatri for dominance (i.e., capitalist motives / logics) emphasises their limits and continuity with platform capitalism models (which shuts down any possibilities for alternative theorisations), while reading them for difference (i.e., as potentially introducing new dynamics) emphasises the specifics of what they have made possible (or not), (which opens up possibilities for alternative theorisations).

What is “new” about reading for difference?

It certainly may be tempting to infer that reading for difference is nothing more than resisting universalisation, localising and nuancing our purviews, and making connections cautiously and tenuously, which are the cornerstones of academic scholarship. So, what exactly is “new” or “different” from what we already know and practice? For me, the key difference is in learning how to re-orient these norms towards deliberately cultivating capacities to “open-up” rather than “shut- down”. It is about asking theory to not only extend knowledge by confirming what we already know, but to provide us a space for constructing alternatives and possibilities. A big part of this comes from challenging our attachment to strong theories of power, which makes any (other) act of “taking back power” appear facetious, ill-conceived, or unviable.

To be clear, we cannot or need not “think” away Big-tech power, but we can un-think Big-tech determinism (Caplan, et.al. 2020). We can attend to the particularities of Big-tech’s stated power (Woods, 2025), while also being reflexive of how our critiques of such power could inadvertently play into the hype (Burgess, 2023). We can do more to theorise the everyday world of technologies and the people engaging with them (Burgess, 2023). Beyond the AI winters and summers of Silicon Valley, capital scarce and localised AI visions set-in real-world problems are driving a multitude of modest innovations across India, Latin America and Africa (Drage, 2025). By attending to the specifics, peculiarities and particularities of these innovations, we can develop new vocabularies and concepts to articulate the landscapes of possibilities. And we can claim and legitimise the space for envisioning what I call “plural digitalities” – i.e., spaces of experimentation and failures, of co-options and transgressions, of cautious connections and learnings. At a time when our “dominant” readings of power infuse the world with a certainty and sufficiency of the alternatives that exist, a “different” reading of power could inspire both the courage and comfort to embrace the contingency and indeterminacy of the paths that will shape our digital futures, and the voices involved in shaping it.

References

Aravind, I. (2025, May 2). Rest of World. Retrieved from Rest of World: https://restofworld.org/2025/uber-ola-copy-india-zero-commission-ride-hailing-app/

BBC. (20016, 8 Feb). India blocks Zuckerberg’s free net app, BBC.

Burgess, J. (2023). Everyday data cultures: beyond Big Critique and the technological sublime. AI & Society, 38(3), 1243-1244.

Caplan, R., Clark, M., & Partin, W. (2020, Oct 13). Against Platform Determinism: A Critical Orientation. Data and Society.

Chaudhary, R. (2025, April 2). How Namma Yatri, Rapido are disrupting the disruptors. The Ken.

Ganapathy, A. (2025, February 19). Can critique encompass hope and optimism. ICTs for Development.

Gibson-Graham, J. K. (2006). A Postcapitalist Politics (NED – New edition ed.). University of Minnesota Press.

Gibson-Graham, J. K. (2008). Diverse economies: performative practices for other worlds. Progress in human geography, 32(5), 613-632.

Drage, E, (2025, August 2). Enough of the billionaires and their big tech. ‘Frugal tech’ will build us all a better world. The Guardian.

Hendrikse, R., Adriaans, I., Klinge, T. J., & Fernandez, R. (2022). The big techification of everything. Science as culture, 31(1), 59-71.

Kapur, A. (2025, June 27). AGI v/s AAI Grassroots Ingenuity and Frugal Innovation Will Shape the Future. Tech Policy Press.

Namma Yatri. (n.d.). Retrieved from https://nammayatri.in/

Namma Yatri. (n.d.). Retrieved from https://nammayatri.in/open?source=in-app&cc=

Nothias, T. (2020). Access granted: Facebook’s free basics in Africa. Media, Culture & Society, 42(3), 329-348.

Pahwa, N. (2015, September 25). A change of name to Free Basics doesn’t make Facebook’s Zero-Rating service neutral. Media Nama.

Parsheera, S. (2022). India’s Policy Responses to Big Tech: And an Eye on the Rise of ‘Alt Big Tech’ Special Issue on FinTech. Indian Journal of Law and Technology, 18(1), 33-65.

Rikap, C. (2024). A fit-for-purpose platform research agenda for a broken world. Platforms & Society, 1, 29768624241263951.

Woods, D. (2025). The Illusion of Power: Unpacking the Vulnerabilities of Big Tech in the Face of Sovereign States. Chinese Political Science Review, 10(2), 149-177.

Digital Transformation as Flow: Why “Lines of Action” matters for ICT4D

30 January 202631 January 2026 ~ nicholsonbrian ~ Leave a comment

Brian Nicholson

Much of what we read about digital transformation treats it as a journey with a map: assess, design, implement, optimize. Whether framed as “roadmaps,” “maturity models,” or “capability building,” the underlying assumption is remarkably stable, digital transformation unfolds as a sequence of planned, managerially guided steps.

But what if this way of thinking fundamentally misses how digital transformation actually happens, especially in resource-constrained contexts?

Our recent study of a UK small accounting firm’s digital transformation challenges this linear logic by reporting on the lived experience of digital transformation theorised by combining bricolage with a flow-oriented perspective. The flow concepts are inspired by the anthropologist Tim Ingold as applied recently by IS scholars (e.g., Baygi et al., 2021), bricolage is established in our discipline from the work of Claudio Ciborra.

Instead of stages and plans, our paper follows lines of action: the evolving, intersecting trajectories of people, technologies, vendors, regulations, breakdowns, and opportunities. Digital strategy, in this view, is not planned and executed, it emerges.

From plans and projects to lines of action

The case follows our participant observation of how the digital strategy team (including the authors) navigated multiple digital initiatives,: automating bank confirmations, improving client data exchange, and rethinking audit workpapers. At no point did these unfold as a clean sequence. Instead, they were shaped by:

vendors releasing (or withdrawing) features,
vendors growing big fast and winning new much larger clients meaning lower priority given to us
regulators legitimising or problematising particular tools,
infrastructure failures (including a flooding incident),
shifting internal attentions and frustrations, including a merger

A flow perspective reframes these not as “external factors” impacting a core strategy, but as multiple lines of action moving at different speeds and rhythms. Vendor roadmaps, regulatory cycles, legacy systems, and everyday audit work all flowed on their own trajectories. We explain how digital transformation occurred when these lines corresponded when they crossed in ways that created openings for action.

Crucially, this brings time back into the foreground — not chronological project time, but kairotic time: moments that matter. The flooding of an office, the announcement of a new vendor partnership, or the withdrawal of promised customisation became pivotal not because they were planned milestones, but because they created transitional moments in which new combinations suddenly made sense.

Bricolage in flow: making do as “ongoing becoming”

Our paper combines flow with bricolage research that shows how organisations “make do” by recombining available resources. What the flow lens adds is a shift from bricolage as episodic problem-solving to bricolage as ongoing becoming.

Resources were not simply “at hand.” They were themselves moving: software packages evolving, APIs appearing and disappearing, regulatory expectations shifting, internal competences sedimenting. What counted as usable today could become obsolete tomorrow and then suddenly relevant again.

Seen this way, digital transformation is not about closing a gap between “current” and “future” states. It is about wayfaring: continuously adjusting direction while already in motion. Strategy becomes less about control and more about attentional skill noticing when lines of action begin to converge, and having the organisational sensitivity to act in those moments.

This way of thinking has profound implications beyond UK based professional service firms especially for ICT4D

ICT4D contexts are, almost by definition, shaped by:

chronic resource constraints,
dependence on external platforms, donors, and vendors,
unstable infrastructures,
layered regulatory and institutional pressures,
and deep entanglements between technology, livelihoods, and everyday practice.

Yet ICT4D interventions are still frequently framed through project imaginaries: pilot → rollout → scale; design → implement → evaluate. These imaginaries struggle to capture what actually unfolds on the ground: patchwork infrastructures, repurposed technologies, shifting workarounds, and constant negotiation between local practices and external systems.

A flow-oriented, lines-of-action perspective invites a different sensibility.

Instead of asking, “Was the system successfully implemented?” we might ask:

What lines of action were already in motion before the technology arrived?
Which new flows did the intervention introduce (data, funding, standards, expertise)?
Where did these lines align, interfere, or fall out of sync?
Which kairotic moments enabled local actors to reconfigure technologies in meaningful ways?

In many ICT4D settings, digital innovation happens precisely through bricolage in flow: health workers combining WhatsApp with paper records; farmers aligning platform updates with seasonal cycles; NGOs adapting donor systems to local reporting rhythms. These are not deviations from the “real” transformation they are the transformation.

Rethinking impact: from outcomes to correspondences

A lines-of-action view also reframes how we think about success and impact in ICT4D. Instead of measuring progress against predefined outcomes, it directs attention to correspondences: how technological, institutional, and everyday lines of action begin to move together (or fail to).

This matters because many ICT4D failures are not technical breakdowns, but temporal and relational misalignments systems introduced before regulatory legitimacy exists, platforms evolving faster than local skills, or donor cycles outpacing community rhythms.

From a flow perspective, the challenge is not only to design better tools, but to cultivate the capacities to:

sense emerging transitional moments,
hold open multiple possible futures,
and work creatively with partial, shifting resources.

Towards a flow-sensitive ICT4D

What this study ultimately contributes is not another framework to “manage” digital transformation, but a vocabulary for recognising and staying with its movement. For ICT4D scholars and practitioners, this means taking seriously that technologies do not enter stable contexts, they enter flows.

In resource-constrained environments whether SME audit firms or rural health systems digital futures are not simply and unproblematically rolled out. They are grown, spliced, resisted, rerouted, and occasionally accelerated.

We are not suggesting anyone abandons structured project management, project plans, Gannt charts and change management techniques, all these techniques remain important. The lessons of the case are that skill in planning the route is complemented by learning how to navigate, as wayfarers within the flows.

Reference

Baygi, R.M., Introna, L.D. and Hultin, L., 2021. Everything flows: Studying continuous socio-technological transformation in a fluid and dynamic digital world. MIS quarterly, 45(1), pp.423-452.

Understanding the Politics of Practice in Development Projects

12 January 2026 ~ Richard Heeks ~ Leave a comment

Diagram of local and global network analysis of a development project

In development, we talk a lot about policy, strategy and results. We talk less about the everyday practice of project implementation. But this messy, political reality is what actually determines whether development interventions succeed or fail.

In our paper, Understanding Development Project Implementation: An Actor-Network Perspective, Carolyne Stanforth and I revisit this neglected space by applying a lens that is surprisingly under-used in development studies: actor-network theory (ANT).

Why implementation practice still matters

Despite the shift towards governance, institutions and high-level reform – and notwithstanding the current aid cutbacks – development projects remain the core mechanism for turning development intentions into development impacts. Yet between intention and impact lies what David Mosse once called the “black box” of project practice.

That black box remains largely unopened. Analyses tend to rely either on linear management models or on binary judgements of success/failure. Both approaches miss the political, organisational and material dynamics that actually shape project trajectories.

The value of actor-network theory

ANT provides one way to grapple with this complexity. Rather than starting from structures or individual decisions, ANT starts from networks: shifting associations of actors (human and non-human) whose alignments must constantly be built and rebuilt.

In the paper, we draw particularly on Law and Callon’s notion of local and global networks. The global network provides the resources and legitimacy for a project; the local network does the work of implementation. Successful projects require both of these networks to be mobilised and, critically, for the project to become an obligatory point of passage between them.

A Sri Lankan reform project as example

We illustrate this through a detailed case: an Asian Development Bank-funded public expenditure reform project in Sri Lanka. Initially, the global network of senior ministry officials and ADB staff aligned behind an ambitious, technology-led reform agenda. But the local network never mobilised. Existing financial systems, long-standing departmental interests, and election-driven political churn produced resistance. The project’s first strategy thus failed because it did not become the shared pathway through which actors saw their interests being met.

Only after a redesign – anchored in a new LogFrame, a new project director, and a more incremental approach – did the networks stabilise. Various deliverables followed, including ministry-wide digital communications, enhancements to existing financial systems, and the Integrated Budget System that ultimately became the visible marker of “success”. As represented in the network analysis diagram below, project progress mirrored changes in the networks, not changes in formal plans.

Diagram of local & global network analysis of a development project

Rethinking power and politics in development management

Perhaps the most useful insight from viewing projects through the ANT lens is about power. The ADB held formal authority and financial leverage but these capacities did not deliver the initial design. Power was enacted only when other actors were successfully enrolled: when associative, not capacitive, power took effect. This reminds us that project management in development is inherently political. It is less about enforcing compliance, more about persuading, enrolling and improvising within complex actor-networks.

Why this matters

By opening up the black box of implementation, ANT helps us understand how development projects really unfold. It highlights the dynamic interplay of actors, technologies, documents and interests. And it provides a language for analysing project trajectories that goes beyond the simple notions of either success or failure. For practitioners and researchers, this perspective offers a richer, more realistic understanding of what it takes to make development projects work in practice.

For more details on applying ANT for analysis of development projects, take a look at the full paper: https://research.manchester.ac.uk/en/publications/understanding-development-project-implementation-an-actor-network/

Rethinking Educational Strategies For Rural India – The Case For Digital Learning Room Project (DLRP)

16 December 2025 ~ Dr Chandrani Maitra ~ Leave a comment

Background and Introduction

A large portion of India’s population still lives in rural villages, many of whom are children. Yet, their educational environment faces numerous complex challenges. This leads to the growth of an underserved community, increasing social inequality. The key to addressing this gap is modernising the educational system to provide children with access to contemporary teaching methods. Digital education has become a powerful tool in children’s learning today. It should be engaging, interactive, and dynamic, fostering comprehensive learning experiences.

The Centre for Social Change and Development (CSCD), a UK-registered charity, is committed to empowering people through knowledge sharing. Its flagship project, Digital Learning Room, involves remote education for young children separated by social and technological barriers (https://www.cscad.org.uk/). This mid-term report evaluates the effects of introducing digital resources in villages within underserved communities.

Methodology and Execution

This initiative by CSCD involved seven interventions across two schools situated 267 km apart (Sargachi and Raidighi), serving children aged 4–7 from socioeconomically disadvantaged communities in the Gangetic delta. The schools, with 30 and 31 students respectively, adopted a blended learning approach on alternate Sundays (10 am – 12 PM), replacing traditional chalk-and-blackboard methods with a SMART TV, a laptop, and smartphones, with support from a lead teacher and an assistant teacher.

The lead teacher conducted sessions via Google Meet, connecting both classrooms remotely. The curriculum—including stories, songs, rhymes, math, and literacy activities—was delivered synchronously to both groups. Assistant teachers supported in-class activities, managed logistics, and communicated with parents. A 15-minute break followed each 45-minute session to keep children engaged.

Before starting, the CSCD coordinator held meetings with teachers and assistant teachers, which involved identifying digital resources, preparing teaching materials and curriculum, and consulting with other staff to share risk management strategies, ensure the smooth running of classes, and so on. Additional volunteers were on standby to take over if needed, with lesson plans designed to be flexible for quick adaptation.

On the day of teaching, the teacher briefed the class about the schedule, noting the distance between groups. Each group was led by an assistant responsible for seating, distributing materials, providing worksheets, and handling needs like fetching water or contacting the local coordinator if someone fell ill. If personal needs arose, the assistant communicated with the CSCD coordinator via WhatsApp, who supervised the project. Parents stayed informed through mobile updates from the teacher about activities, homework, and cancellations.

During instruction, the teacher assigned tasks to each group, with assistant teachers carrying them out. Contingency plans for power outages included having digital resources available offline. For example, during an unexpected power outage at Raidighi, the teacher showed a video on his smartphone to one group, while others engaged in activities such as building with blocks, colouring, or decorating houses. Once power was back, groups shared their experiences.

Key outcomes assessed

Key Outcomes Assessed

1.⁠ ⁠Equity in Education: Geographical barriers were bridged, offering high-quality instruction to remote learners.

2.⁠ ⁠Teacher Capacity Building: Co-teachers evolved into complementary roles, leveraging technology for seamless execution.

3.⁠ ⁠Resilience: The model proved adaptable to disruptions, ensuring continuity.

4.⁠ ⁠Community Engagement: Parents and educators embraced the approach, advocating its sustainability.

Impact on Learners

The digital approach created an interactive environment where children actively engaged with audio-visual content, peer interactions, and creative activities such as drawing and crafting. The inclusion of music, drama, and storytelling encouraged greater participation. All 61 children attended consistently and showed enthusiasm for ongoing learning.

Parental Perspectives

Parents valued transparent communication and the opportunity to observe their children’s progress. Coordinators stayed in contact through phone calls and shared performance records.

Teacher Perspectives

Teachers and assistants felt empowered through collaborative lesson planning, real-time support, and access to digital resources, which were especially helpful in managing classes during emergencies such as storms.

Figure: Activities conducted (all photographs obtained with consent)

Discussion

This intervention highlights the potential of digital tools to transform early education in resource-limited settings, promoting inclusive and joyful learning environments. This discussion explores how digital learning empowers rural children, improves their learning outcomes, and supports societal development.

The transformative role of digital learning in rural indian education: bridging the educational divide

In rural India, limited skilled teachers, poor infrastructure, and scarce learning resources create significant barriers to quality education. These areas often lack trained educators and teaching aids, resulting in disengaged students and high dropout rates. Nonetheless, digital learning rooms have the potential to transform early childhood education in these underserved regions. By combining technology with teaching methods, these initiatives can fill educational gaps and make learning more engaging and inclusive.

Digital learning rooms, equipped with SMART TVs, laptops, and internet access, facilitate virtual interactions with expert educators. For example, one teacher can instruct multiple classrooms simultaneously via platforms like Google Meet, maintaining consistent teaching quality. Interactive content—such as videos, animations, and gamified modules—engages children and makes complex concepts easier and more enjoyable to understand. This strategy not only boosts academic achievement but also helps narrow the educational gap between urban and rural areas.

Fostering holistic development

Digital tools expand learning beyond traditional textbooks. Children can visit virtual museums, join storytelling sessions, and participate in activities such as digital art creation. These experiences foster creativity, critical thinking, and communication skills, essential for overall development. Moreover, access to global content helps broaden their horizons, sparks curiosity, and fuels ambitions. Even a child in a remote village now has the chance to learn about the solar system, coding, or environmental conservation—opportunities once limited to urban elites.

Empowering teachers and communities

Digital learning supports educators with accessible resources and training. Assistant teachers in rural schools, who are often underprepared, get real-time support, boosting their confidence and teaching skills. Parents also become active participants in their children’s education by staying informed about curriculum updates and monitoring their children’s progress. This collaborative ecosystem fosters community involvement, driving ongoing engagement.

Overcoming socio-economic barriers

Digital education helps marginalised communities overcome geographical and economic barriers. It connects children to the wider world, allowing them to surpass conventional limits. According to the UNESCO Institute for Statistics, access to digital technology can boost literacy rates and lessen gender gaps in education. Providing children with 21st-century skills through digital learning equips them to engage in the global economy.

Challenges and the Way Forward

Digital learning has great potential but encounters challenges such as unreliable internet, power outages, and the digital divide. To address these issues, solutions like solar-powered infrastructure, offline resources, and teacher training programs are essential. It is crucial for governments and NGOs to work together to expand these efforts, ensuring that vulnerable populations are not left behind.

Conclusion

Digital learning is not merely a luxury but a fundamental necessity for the children in rural India. Initiatives in digital education aimed at rural regions effectively address significant gaps in access and quality of education. By integrating cutting-edge technology with pedagogical methods, these programs foster greater student engagement, enhance learning outcomes, and empower marginalised communities. In promoting engagement, equity, and innovation, such initiatives contribute to the development of a more educated and equitable society.

Nonetheless, the sustainable implementation of these programs necessitates comprehensive infrastructure development, extensive teacher training, and inclusive policy frameworks. As the global community advances in digital transformation, rural India must strategically harness this opportunity to realise its demographic potential. This treatise asserts that digital education serves as a pivotal driver for equity and socio-economic development, and calls upon stakeholders to focus on scalable, contextually appropriate solutions. The future progress of India fundamentally depends on bridging the digital divide today.

Resisting the Narrowing of Education: Emerging Critical AI Competencies Among Students in Bangladesh and the UK

11 December 202512 December 2025 ~ taslimaivy ~ Leave a comment

Dr Taslima Ivy, Lecturer, Manchester Institute of Education, University of Manchester

Dr Martyn Edwards, Lecturer, Manchester Institute of Education, University of Manchester

Many discussions about Artificial Intelligence in education position teachers and students as passive recipients of technological change, while corporations, policymakers, and technical experts define what counts as effective AI use (Miao and Shihohira, 2024, p.12). Velander et al. (2024) argue that such approaches render individuals “powerless to change or disrupt this future,” signalling a deeper mechanism of disempowerment. Selwyn (2022) adds that the most significant digital technologies of the 2020s are often “used on” people rather than “used by” them. Building on this, Selwyn (2024) shows how, in education, classrooms are increasingly redesigned to fit what data systems can measure, producing what he calls ‘recursive narrowing,’ where technology, rather than teachers and learners, decides what counts as learning.

Can Critical AI competencies offer an alternative? Darvin (2025) argues these competencies extend beyond mere functionalistic skills, allowing learners to interrogate how power, ideology and inequality are reproduced within digital spaces. Dindler et al. (2020) emphasises that young people need more than coding skills; they require opportunities to examine how technologies are imbued with values and how they might act to change them. Such perspectives resonate with Freirean pedagogy (2000) which centres upon students lived experience, agency and capacity to co-create knowledge.

To explore this potential practice, we developed the AI in Education seminar series and AI Literacy Hackathon—an initiative between the University of Manchester and three Bangladeshi universities (University of Dhaka, University of Liberal Arts Bangladesh, Noakhali Science and Technology University), supported by a2i, Government of Bangladesh. The aim was not to teach AI as a set of neutral skills but to create a space for inquiry, reflection, collaboration and creation through four interlinked phases:

Collective Inquiry: Eight seminars brought together students, educators, policymakers, and ed-tech professionals as a community of co-learners to examine how values, power, and equity shape the design and use of AI and how these sociotechnical systems impact individuals and communities.
Critical Reflection: Reflective tasks, facilitated through Padlet, allowed students to further analyse issues raised in the seminars and articulate concerns they developed after the live sessions.
Collaborative Outputs: In a final hackathon, students collaborated to create AI literacy materials (comic strips, podcasts, reels etc.) designed for real-world use and spark broader conversations about AI
Public Recognition: Student outputs were celebrated via a final presentation and certificate ceremony.

Figure 1: Student-created comic excerpts examining AI in education

The student-created comic strips below are samples of the artefacts produced through this process, offering a deeper window into how learners engaged with critical questions around AI in education.

The final materials demonstrate students’ critical orientations toward technology, identity, and power confirming that bottom-up participation can foster rich forms of criticality. For instance, in the podcasts students voiced concerns that many languages, including their grandmother’s, are absent from AI datasets, provoking fears of linguistic erasure. Some described AI’s constant corrections as a form of indirect colonisation of American norms. One comic strip captured cultural mismatch, noting that ‘AI doesn’t fully understand our culture.’ Students also uncovered systemic issues. For instance, Bengali participants described AI as a ‘non-judgmental friend’ within environments where asking questions feels unsafe thereby challenging traditional hierarchical student/teacher dynamics, leading another group to attest ‘the problem is within us, not AI.’ Others highlighted digital inequities like limited AI use due to lack of devices or internet access.

In a global context dominated by top-down AI agendas, these perspectives reveal issues often missing from mainstream policy conversations. At the same time, many of the proposed solutions remained early stage, for example, creating more data rather than imagining alternatives. This highlights that critical AI competencies develop gradually through iterative cycles of inquiry, design, and feedback. Taken together, the hackathon artefacts surface a set of generative questions that reposition AI in education not as a solution to be optimised, but as a sociotechnical issue to be critically explored:

How are students and teachers currently using AI in practice, and what needs are these uses responding to?
What systemic failures are students attempting to solve through AI?
How do AI systems align with or colonise existing cultural/epistemological values in a context?
What kinds of participatory, iterative structures are needed for educator/learners to co-shape AI practices and build educational futures?

If you are interested in co-designing critical AI competencies with your students or communities or joining our next seminar series and hackathon starting in February 2026, please feel free to get in touch by emailing Ivy (taslima.ivy@manchester.ac.uk ) or Martyn (martyn.kj.edwards@manchester.ac.uk).

References:

Darvin, R. (2025). The need for critical digital literacies in generative AI-mediated L2 writing. Journal of Second Language Writing, 67, 101186.

Dindler, C., Smith, R., & Iversen, O. S. (2020). Computational empowerment: participatory design in education. CoDesign, 16(1), 66-80.

Freire, P. (2000). Pedagogy of the oppressed (30th anniversary ed., M. B. Ramos, Trans.). Bloomsbury.

Miao, F., & Shiohira, K. (2024). AI competency framework for students. UNESCO Publishing.

Selwyn, Neil (2022) What should ‘digital literacy’ look like in an age of algorithms and AI? Parenting for a Digital Future (06 Apr 2022). Blog Entry.

Selwyn, N. (2024). On the limits of artificial intelligence (AI) in education. Nordisk tidsskrift for pedagogikk og kritikk, 10(1), 3-14.

Velander, J., Otero, N., & Milrad, M. (2024). What is critical (about) AI literacy? Exploring conceptualizations present in AI literacy discourse. In Framing futures in postdigital education: Critical concepts for data-driven practices (pp. 139-160). Cham: Springer Nature Switzerland.

Beyond Hardware: How AI and Chinese Innovators are Reshaping Digital Accessibility

5 December 2025 ~ Weiyi Zhang ~ Leave a comment

Accessibility is a fundamental concept, ensuring that no one is excluded from participating in society on the basis of a disability. Historically, our efforts focused on removing physical impediments, such as architectural barriers or the need for specialised assistive hardware. However, the rapid advancement of the digital age has introduced complex new information barriers. The challenge is clear: how can accessibility development keep up with this trend? More importantly, can the rise of powerful AI technology be leveraged to bridge the information divide rather than widen it? This post explores these questions by examining the practices of leading Chinese technology firms like OPPO, Vivo, Xiaomi, and HONOR.

The Era of Multimodal Interaction

The mobile phone remains a key portable terminal in our daily lives, and there is a consensus among Chinese tech firms that multimodal interaction is the next frontier. This technology vastly enriches the ways users see, hear, and touch, allowing information to be flexibly translated across different sensory channels to bridge the information divide. Vivo, for instance, believes that intelligent devices act as “external organs” for users. Their feature, “Vivo 看见” (Vivo See), assists users with visual impairments by describing real-world scenes in detail. It can enable a user to confidently visit a flower shop to buy red carnations or, as shown in recent demos, identify exactly how much of a birthday cake has been eaten and where it has been cut.

(Source: Vivo看见)

(Sentences on the picture: “How much of the cake was cut off?” “There is a triangular area at the bottom right corner of the cake that has been cut off, while the rest of the cake is intact.” )

Innovating for Situational and Environmental Barriers

(Source: Google Images)

These innovations extend beyond permanent disabilities to address situational barriers encountered by diverse users. For example, a solitary woman can protect her privacy by using an AI function to reply to calls with a man’s voice, instructing a delivery to “put the takeout outside the door”. Furthermore, new device forms like foldable phones are creating new possibilities: a half-folded device can use the outer screen as a “translator” for displaying information, while the inner screen communicates with the user via lip-reading and gesture recognition. Moreover, by leveraging these multimodal capabilities, more natural multisensory information transfer is being achieved through devices like bone conduction headphones, precise haptic feedback (such as watch vibrations), and holographic projection, enhancing the user experience across various accessibility needs. In the home environment, Xiaomi has established an integrated product series centred on the AI robot “Xiao Ai.” This allows users to interact with furniture and appliances via voice or gesture, providing real-time subtitles for those who cannot hear and scene descriptions for those who cannot see, effectively extending accessibility across the entire house.

(Source: Google Images)

Embedding Inclusion into Business Strategy

The vision of accessible products cannot be realised without embedding it into business strategies. Chinese firms are recognising that inclusion is a driver of value, not just a cost centre. OPPO, for example, is developing a “千人千屏” (A Unique Display for Every User) experience. By leveraging AI to tailor digital interfaces to the needs of every individual, they are aligning with the principle of inclusive design, which is meeting the diverse needs of a wider range of users to increase utility and market appeal. Similarly, HONOR has established a dedicated accessibility R&D group to coordinate cross-departmental efforts and actively invites users to participate in design. Notably, they set up accessibility experience booths at product release conferences to enhance employee awareness of these achievements.

The Ecosystem and the Reality Check

Advancing accessibility requires an industry-wide effort. Vivo has taken significant steps by integrating industry and academia, collaborating with organisations like the China Disabled Persons’ Federation (CDPF), China Association of the Blind (CAB), and universities to gain a deep understanding of users’ needs. In 2024, they freely opened their AI capabilities to developers serving users with disabilities, with daily average calls reaching 2 million, covering many non-Vivo terminal enterprises.

The Other Side of the Coin

However, while these developments are impressive, we must be cautious and look beyond the technological optimism. Are these solutions truly accessible to everyone? Many of these advanced multimodal features rely on the powerful processing capabilities of high-end flagship devices, potentially excluding users in low-resource settings who rely on budget devices. Furthermore, features that constantly analyse the environment via cameras, while helpful for navigation, raise significant questions about data privacy. There is also the risk of AI hallucinations. In a casual setting, a wrong description of a cake is fine. But in a navigational or medical context, an AI hallucination could be dangerous.

Conclusion

The essence of product accessibility lies in empowering people and enabling everyone to equally enjoy the life brought by technology. Chinese tech firms are actively tackling these challenges by focusing on AI-powered terminal innovation, embedding inclusive design into commercial strategy, and building open ecosystems. These efforts promise to promote digital equality, but the journey is far from over. As we celebrate these advancements, we must remain vigilant about the structural challenges of cost and privacy to ensure technology truly bridges the divide.

Acknowledgements

The cases presented in this blog are mainly derived from the roundtable discussion of the 7th Technology Accessibility Development Conference (2025TADC), Beijing, China.

Further reading

DESA (2013) Accessibility and Development: Mainstreaming disability in the post-2015 development agenda. New York: UN-Department of Economic and Social Affairs.
Dritsas, E. et al. (2025) ‘Multimodal Interaction, Interfaces, and Communication: A Survey’, Multimodal Technologies and Interaction, 9(1), p. 6. Available at: https://doi.org/10.3390/mti9010006.
Gilbert, R.M. (2019) Inclusive Design for a Digital World: Designing with Accessibility in Mind. Apress.
Jaeger, P.T. (2022) Disability and the Internet: Confronting a Digital Divide. Boulder: Lynne Rienner Publishers (Disability in Society). Available at: https://doi.org/10.1515/9781626371910.
Makuwira, J. (2022) ‘Disability-inclusive development’, in The Routledge Handbook of Global Development. Routledge.
Persson, H. et al. (2015) ‘Universal design, inclusive design, accessible design, design for all: different concepts—one goal? On the concept of accessibility—historical, methodological and philosophical aspects’, Universal Access in the Information Society, 14(4), pp. 505–526. Available at: https://doi.org/10.1007/s10209-014-0358-z.
Raja, D.S. (2016) Bridging the Disability Divide through Digital Technologies. Background Paper for the 2016 World Development Report: Digital Dividends. World Bank.

Share this:

Share this:

Share this:

An illustrative walkthrough – AI-assisted rapid literature review

What does this mean for research as a role?

Power dynamics – a thorn in AI’s side or can it help us see them?

Is Shumer right? Is AI coming for research too?

Share this:

Share this:

Share this:

Share this:

Share this:

Share this:

Share this: