Category Archives: Education

Topics related to the education sector (usually the tertiary or Higher Education sector).

Building Research Capacity with AI

Over 25 years ago, the “10/90 gap” was used to illustrate the global imbalance in health research. Only 10% of global research benefited the regions where 90% of preventable deaths occurred. Since then, efforts to improve research capacity in low- and middle-income countries (LMICs)—where 90% of avoidable deaths occurred—have made important gains; nonetheless, significant challenges remain. A quarter of a century later, there are still too few well-trained researchers in LMICs, and their research infrastructure and governance are also inadequate. The scope of the problem increased dramatically in 2025 when governments cut North American and European overseas development assistance (ODA, i.e., foreign aid) precipitously. That aid—however inadequate—supported improvements in research capacity.

Traditional approaches to improving research capacity, such as training workshops and degree scholarship programs, have gone some way to address the expertise challenge. However, they fall short because they are not scalable. The relatively recent introduction of massive open online courses (MOOCs), such as TDR/WHO’s MOOCs in implementation research, goes a long way to overcoming that scalability problem—at least in instruction-based learning. Nonetheless, for many LMIC researchers, major bottlenecks remain because of poor or limited access to mentorship, one-off and quick advice, bespoke training, research assistance, and inter- and intra-disciplinary collaboration. The scalability problem can leave them at a persistent disadvantage compared to their high-income country counterparts. Research is not done well from isolation and ignorance.

The rise of large language model artificial intelligence (LLM-AIs) such as ChatGPT, Mistral, Gemini, Claude, and DeepSeek offers an unprecedented opportunity…and some additional risks. LLM-AIs are advanced AI models trained on vast amounts of text data to understand and generate human-like language. They are flexible, multilingual, and always available (24/7), offering researchers in LMICs immediate access to knowledge and assistance. If used correctly, LLMs could revolutionise approaches to building research capacity and democratise access to skills, knowledge, and global scientific discourse. Many online educational providers already integrate LLM-AIs into their instructional pipelines as tutors and coaches.

Unfortunately, LMICs risk further entrenching or increasing the 10/90 gap if they cannot take advantage of the benefits of LLM-AIs.

AI as a game changer

Researchers in resource-limited settings can access an always-on, massively scalable assistant for the first time. By massively scalable, every researcher could have one or more 24/7, decent research assistants for a monthly subscription of less than $20. They offer scalability and flexibility that traditional human research assistants cannot (and should not) match. However, they are not human and may not fully replicate a human research assistant’s nuanced understanding and critical thinking—and they are certainly less fun to have a cup of coffee with. Furthermore, the effectiveness of LLM-AIs depends on the sophistication of the user, the task complexity and the quality of input the user provides.

I read a recent post on LinkedIn by a UCLA professor decrying the inadequacies of LLM-AIs. However, a quick read of the post revealed that the professor had no idea how to engage appropriately with the technology.

Unfortunately, like all research assistants, senior researchers, and professors, LLM-AIs can be wrong. Like all tools, one needs to learn how to use them with sophistication.

In spite of any inadequacies, LLM-AIs can remove barriers to research participation by offering tutoring on complex concepts, assisting with literature reviews and data analysis, and supporting the writing and editing of manuscripts and grant proposals.

Reid Hoffman, the AI entrepreneur, described on a podcast how he used LLM-AIs to learn about complex ideas. He would upload a research paper onto the platform and ask, “Explain this paper as if to a 12-year-old”. Hoffman could then “chat” with the LLM-AI about the paper at that level. Once comfortable with the concepts, he would ask the LLM-AI to “explain this paper as if to a high school senior”. He could use the LLM-AI as a personal tutor by iterating-up in age and sophistication.

Researchers can also use the LLM-AIs to support the preparation of scientific papers. This is happening already because an explosion of generically dull (and sometimes fraudulent) scientific papers is hitting the market. This explosion has delighted the publishing houses and created existential ennui among the researchers. The problem is not the LLM-AIs—it is in their utilisation, and it will take time for the paper production cycle to settle.

While access to many LLMs requires a monthly subscription, some LLM-AIs, like DeepSeek, significantly lower costs and accessibility barriers by distributing “open weights models”. Researchers can download these open weights models freely and put them on personal or university computer infrastructure without paying a monthly subscription. They make AI-powered research assistance viable for most LMIC research settings, and universities and research institutes can potentially lower the costs further.

LLM-AIs allow researchers in LMICs to become less dependent on high-income countries for training and mentorship, shifting the balance towards scientific self-sufficiency. AI-powered tools could accelerate the development of a new generation of LMIC researchers, fostering homegrown expertise and leadership in relevant global science. They are no longer constrained by the curriculum and interests of high-income countries and can develop contextually relevant research expertise.

The Double-Edged Sword

Despite its positive potential, the entry of LLM-AIs into the research world could have significant downsides. Without careful implementation, existing inequalities could be exacerbated rather than alleviated. High-income countries are already harnessing LLM-AIs at scale, integrating them into research institutions, project pipelines, training, and funding systems. LMICs, lacking the same level of investment and infrastructure, risk being left behind—again. The AI revolution could widen the research gap rather than close it, entrenching the divide between well-resourced and under-resourced institutions.

There is also a danger in how researchers use LLM-AIs. They are the cheapest research assistants ever created, which raises a troubling question: will senior researchers begin to rely on AI to replace the need for training junior scientists? Suppose an LLM-AI can summarise the literature, draft proposals, and assist in the analysis. In that case, there is a real risk that senior researchers will neglect mentorship, training and hands-on learning. Instead of empowering a new generation of LMIC researchers, LLM-AIs could be used as a crutch to maintain existing hierarchies. If institutions see the LLM-AIs as a shortcut to productivity rather than an investment in building research capacity, it could stall the development of genuine human expertise.

Compounding these risks, AI is fallible. LLM-AIs can “hallucinate”, generating false information with complete confidence. They always write with confidence. I’ve never seen one write, “I think this is the answer, but I could be wrong”. They can fabricate references, misinterpret scientific data, and reflect biases embedded in their training data. If used uncritically, they could propagate misinformation and skew research findings.

The challenge of bias is not to be underestimated. LLM-AIs are trained on the corpus of material currently available on the web, reflecting all the biases of the web–who creates the content, what content they create, etc.

Furthermore, while tools like DeepSeek reduce cost barriers, commercial AI models still pose a financial challenge. LMIC institutions will need to negotiate sustainable access to AI tools or risk remaining locked out of their benefits—particularly of the leading edge models. The worst outcome would be a scenario where HICs use AI to accelerate their research dominance while LMICs struggle to afford the very tools that could democratise access.

A Strategic Approach

To ensure LLM-AIs build rather than undermine research capacity in LMICs, they must be integrated strategically and equitably. Training researchers and students in AI literacy is paramount. Knowing how to ask the right questions, validate AI outputs, and integrate results into research workflows is essential. This is not a difficult task, but it takes time and effort, like all learning. The LLM-AIs can help with the task—effectively bootstrapping the learning curve.

Rather than replacing traditional research capacity building, LLM-AIs should be embedded into existing frameworks. MOOCs, mentorship programs, and research fellowships should incorporate LLM-AI-based tutoring, iterative feedback, and language support to enhance—not replace—human mentorship. The focus should be on areas where LLM-AI can offer the greatest immediate impact, such as brainstorming, editing, grant writing support, statistical assistance, and multilingual research dissemination.

Institutions in LMICs should also push for local, ethical LLM-AI development that considers regional needs. This push is easier said than done, particularly in a world of fracturing multilateralism. However, appropriately managed, LLM-AI models can be adapted to recognise and integrate local research priorities rather than merely reinforcing an existing scientific discourse. The fact that a research question is of no interest in high-income countries does not mean it is not critically urgent in an LMIC context.

Finally, securing affordable and sustainable access to AI tools will be essential. Governments, universities, and research institutions must lobby for cost-effective AI licensing models or explore open-source alternatives to prevent another digital divide. Disunited lobbying efforts are weak, but together, across national boundaries, they could have significant power.

An Equity Tipping Point

The LLM-AI revolution is a key juncture for building research capacity in LMICs. Harnessed correctly, LLM-AIs could break down long-standing barriers to participation in science, allowing LMIC researchers to compete on (a more) equal footing. The rise of models like DeepSeek suggests a future where AI is not necessarily a privilege of the few but a democratised resource for the many.

Fair access will not happen automatically. Without deliberate, ethical, and strategic intervention, LLM-AIs could reinforce existing research hierarchies. The key to harvesting the benefits of the technology lies in training researchers, integrating LLM-AIs into programs to build research capacity and securing equitable access to the tools. Done well, LLM-AIs could be a transformative force, not just in scaling research capacity but in redefining who gets to lead global scientific discovery.

LLM-AIs offer an enormous opportunity. They could either empower LMIC researchers to chart their own scientific futures, or they could become another tool to push them further behind.


Acknowledgment: This blog builds upon insights from a draft concept note developed by me (Daniel D. Reidpath), Lucas Sempe, and Luciana Brondi from the Institute for Global Health and Development (Queen Margaret University, Edinburgh), and Anna Thorson from the TDR Research Capacity Strengthening Unit (WHO, Geneva). Our work on AI-driven research capacity strengthening in LMICs informed much of the discussion presented here.

The original draft concept note is accessible here.

Leaders can be bullies too.

Leaders can be bullies too. And their poor behaviour will infect the whole organisation.

When I hear the word “bully“, even at work, I inevitably recall the schoolyard bullies of my youth. Often with a clique of sycophants, they were the nasty kids who tried to intimidate others. Their gangs were not deeply committed to being mean. They were committed to survival. Better, they reasoned, to support a thug than get sand kicked in their faces. Or worse, become the butt-end of cruel taunts about bad haircuts.

Unfortunately, we do not leave the bullies behind when we leave the playground. Bullies grow up and find their niche in adult life. The ease with which they establish themselves in an organisation—think parasitic wasp, not butterfly—signals the workplace’s tolerance for bad behaviour

In an organisation with a strong supportive culture, managers deal with bullying swiftly and seriously. Minor incidents are treated as teachable moments. At low levels, the strategy may be as simple as one colleague being empowered to stand up for another—to make it known there is a line in the sand. At higher levels, when bad behaviour escalates, complaints about bullying are heard, taken seriously, and investigated rather than diverted and buried.

In one organisation I worked for, the Chief Executive Officer (CEO) was a well-known, old-school playground bully without the finesse one might expect from a modern leader.

One day, he wandered into my office. He didn’t like my research group’s strategy and wanted to tell me so. Dropping into a chair without greeting or invitation, he rocked back and started into me. I held my position. He became angrier and raised his voice. His reputation for shouting preceded him, and I was prepared. I had decided to match him decibel for decibel. He became louder; I became louder. 

He quickly realised that we were shouting at each other and began to drop the volume. I followed suit. For about 10 minutes, the loudness of the conversation rose and fell. At the end of it, he smiled at me, said, “good chat”, rocked himself out of the chair and left. We had not agreed, but we had reached a rapprochement, and he left me to manage my own team.

I would not recommend my strategy even though it worked at the time. It can be extremely frightening to have a large adult male shout at you. It is also precisely why they do it. Unless you can cope with the aggressiveness of the interaction (and frankly, why should you?), shouting back is not going to work. It’s also unprofessional and fails to address the more significant structural issue. 

Bullying was a regular tactic in my boss’s amentarium, and I achieved a temporary, personal solution that left others exposed. Because no one had ever managed his behaviour, his experience was that shouting worked. It was rewarded by compliance, and compliance was what he wanted.

Much of the leadership literature is about the qualities that one requires to “bring people along”, sell a vision, encourage engagement, (re-)align activities, and gather support for the (new) organisational strategy. The CEO short-circuited that messy business by bullying staff. Instead of intelligent workers, he wanted compliant widgets. The tactic, however, is stupid and lazy. Leaders who adopt it will lose one of their greatest assets. Disempowering staff reduces an organisation’s human capital. The short term win of reluctant compliance is offset by a deterioration of morale, the loss of good employees, and an absence of fresh perspectives. Organisations that accept bullying in leadership tacitly agree to become weaker organisations

Bullying is also a quickly learned behaviour that obviates the need for senior staff to hone their leadership skills. If at first you don’t succeed, shout louder. Others learn the strategy, and it becomes an existential danger for the organisation.

Unfortunately, bullies in leadership are often not ranting, physical thugs and they don’t wear convenient labels. “BEWARE, BULLY!!!”. They have more polished and sophisticated tacticsThe techniques can be pretty subtle and their true nature is often concealed from those who are not the targets. 

When the most senior person in the organisation is a bully, who then will take action? The organisation’s Board or equivalent should step in, but this is easier said than done. The bullied staff member needs to know how to raise their concerns to the Board, and the Board needs to have the willingness to listen and act.

For a bullied staff member to complain, they have to believe it will make a difference. Unfortunately, complaining is often the employment equivalent of stretching your neck out on the chopping block. The victim needs to trust the process, and many organisations provide no basis for that trust. For managing bullies in leadership, the process should be well known, straightforward, and direct to the Board. It never entered my head to complain about my former CEO. I thought it was my problem, and I did not know of any internal processes, let alone a route to the Board. There are also, almost certainly, gender dimensions to who is bullied, how they manage it, and how seriously they are taken.

To manage bullying complaints about leaders, Board members need to be informed, engaged, and empowered to take the complaints seriously. “The Board has an absolute and unmistakable obligation to exercise oversight of workforce culture“. For NGOs, not-for-profits and other non-commercial Boards, membership is often voluntary or unremunerated. Such part-time, “not too serious” Boards can be particularly vulnerable to Directors’ and Trustees’ ignorance and lack of training. There are also disincentives for Boards to take bullying complaints seriously about senior leadership.

The CEO is usually a member of the Board and a colleague of the rest of the members. Some of the Board members will have been nominated by the CEO. Others may have been a part of the CEO’s selection process. When the CEO nominates a person to the Board, the nominee’s sense of loyalty can cloud their judgment about the CEO’s wrong-doing. After all, if the CEO nominated me, she must be OK because I’m great. When the CEO is found wanting, there may be a real sense of failure or a loss of face by Board members involved in the appointment. If a CEO is a bully, clients and the senior leadership team may question the Board’s competence and seek a review of the due diligence processes, with all the attendant embarrassment that can flow from that. All these impediments encourage Boards to obfuscate.

A quick internal process in the guise of swift action is a short-term (wrong-headed) solution to complaints about senior leadership bullying. The result is a superficial examination of the complaint that gives the Board comfort. It allows for a peremptory dismissal of the complaint and avoids embarrassment or culpability. It is easy to imagine, for instance, excusing bullying as a matter of “management style” rather than seeing it for what it is. This is wrong. There is nothing stylish about a bully. Unfortunately (or perhaps, fortunately), superficial processes for managing leadership misconduct have a nasty habit of coming back to bite an organisation. 

A better approach, which carries a higher initial cost, is to engage an external, independent party. Let them investigate the complaint. It demonstrates the matter is being taken seriously, managed impartially, and led by the evidence. It also sets a loud, zero-tolerance tone within the organisation, setting or reinforcing the organisational culture.

If there are any concerns that bullying may be ongoing, administrative leave for the CEO (without prejudice) can be applied while an investigation is conducted. An excellent example of this was the suspension of the newly appointed Director of SOAS following a complaint of racism. The suspension occurred within months of his appointment, and following an investigation, he was cleared and reinstated. Any initial embarrassment that may have been felt is washed away by sound processes.

Unfortunately, the entire premise of this piece rests on two things. First, staff must be prepared and able to raise concerns about bullying by those in leadership. Second, the Board must be trained, competent and serious about managing it. Pretty words are not enough. 

Staff realities are such that it can be better to suffer in silence or leave the organisation. I have known numerous staff of various organisations who chose to go rather than complain about their toxic workplace. Until you have witnessed the pyrotechnic career collapse of those who complained and were not heard, it is sometimes difficult to understand the reluctance. 

No one wants to join the ranks of the pilloried complainers. The received wisdom is to “slip away” or “put up with it”. If Boards are not prepared to hold CEOs accountable, “slip away” is sound advice—tragic and indicting, but sound.

Research brain drain from the global south

The Director of the School of Oriental and African Studies (SOAS) in London, Dr Adam Habib, recently argued that universities in the global north are taking the best and the brightest from the global south and failing to return them.

360info asked me to reflect on this for a special issue on the education brain drain, and write about it from the perspective of research in the global south. What I wrote builds on previous ideas I’ve published and blogged about around the idea of “trickle down science” and decolonising research. This is an edited version of the 360info article.


The indigenous Bajau Laut of southeast Asia live a nomadic existence at sea. They have lived on houseboats for more than 1,000 years, free-diving for marine resources to sustain themselves. Research on the human genetic changes that allowed the Bajau Laut to adapt to this life at sea was published in 2019 in Cell. All but one of the article’s authors came from developed economies. The one Indonesian researcher had no relevant disciplinary background and appeared to be logistical support. The Indonesian government saw the study as exploitative and legislated to restrict overseas researchers from fly-in, fly-out, “grab the data and run” research. 

It’s an example of a common problem: the world’s poorest economies suffer health and development deficits that require research, but they are least likely to do research. When they do research with developed economy collaborators, it is often not the most relevant research to the developed economy.

The highest-income economies graduate the most PhDs per capita — the principal qualification for researchers — whilst the poorest economies graduate the least. The current stop-gap solution, critiqued by Dr Habib, is for developing economies to send their best and brightest students away to overseas PhD programs, often in developed economies. But the PhD experience in developed economies is usually geared towards research training involving sophisticated techniques and equipment unavailable at home. The student cannot replicate the research environment when they return to their home institutions and fall into an intellectual suzerainty. 

A supplementary approach to improving research capacity is through research collaborations. Many developed economy researchers enjoy the opportunity to collaborate with developing economy researchers. The developed economy researchers offer much-needed injections of capital and equipment; they can also provide experience using the latest collection techniques or analytic methods. Through the collaborations, developing economy researchers grow their skills and their networks. They are also much more likely to become authors of well-cited journal articles, which improves their international standing. 

However, significant concerns have been raised recently about the nature of the research collaborations between developed and developing economies. The concerns pivot on whether the relationship is exploitative. Are the collaborators from developing economies equal partners in the research, or are they logistical support, as in the case of the Bajau Laut study? Improving research capacity in developing economies needs to be realistic about the challenges and the structural deficits. There needs to be mutual respect. And it needs to be resilient to foreseeable and unforeseeable shocks. 

Around 10-years ago, the Wellcome Trust funded a project to establish a virtual institute for interdisciplinary research of infectious diseases of poverty in four countries (five institutions) in West Africa. Two developed economy institutions provided support. Nigeria and Mali had Boko Haram insurgencies during the project, and Côte d’Ivoire had a coup. Unfortunately, these external shocks are not atypical examples of the challenges of research capacity strengthening.

Political upheaval notwithstanding, the North-South-South (NSS) approach taken in developing the virtual institute was promising. The project networked developing economy institutions with some developed economy institutions, and it focused on the institutes, not on individual researcher capacity—which is easily lost. It is more holistic and looks to the development of infrastructure, governance, and human capital. Because the approach is based on a multilateral partnership, there are opportunities for mutual support within and between institutions and individual researchers. Governance developments in one institution can be replicated and adapted in another. Depending on the nature of the research, infrastructure can also be shared, such as cloud computing and gene sequencers.

The Norwegian government uses this approach, as does the World Health Organization, albeit in a slightly different form. The NSS approach also stands in marked contrast to supporting one-off projects or funding individual research degrees. The NSS PhD training is based in the developing economy institutions with support from the developed economy institutions in the network, including support from supervisors in the developing economies institutions. The approach simultaneously builds the developing economies’ supervisory capacity and decreases the likelihood of brain drain. The research is also driven by the relevance of the research to the developing economies and utilises technology that is available. 

It is not possible to mandate mutual respect. Developed economy institutions that have been successful over the past half-century in the traditional engagement models — “send your brightest and we will train them”, or “here’s some money, send the data” — may find changes in the status quo unappealing. However, there is no doubt that the NSS approach requires a different mindset, particularly in the institutions of the global north. The research capacity needs of the global south are enormous. The traditional approaches can not meet the needs because they do not scale. New global north institutional players will be needed, and they won’t have the baggage of past practice to weigh them down.


The original article was published under Creative Commons by 360info™. This is an edited version.

Local causation and implementation science

If you want to move a successful intervention from here (where it was first identified) to there (a plurality of new settings), spend your time understanding the context of the intervention. Understand the context of success. Implementation Science—the science of moving successful interventions from here to there—assumes a real (in the world effect) that can be generalised to new settings. In our latest (open access) article, recently published in Social Science and Medicine, we re-imagine that presumption.

As researchers and development specialists, we are taught to focus on causes as singular things: A causes B. Intervention A reduces infant mortality (B1), increases crop yields (B2), keeps girls in school longer (B3), or…. When we discover the new intervention that will improve the lives of the many, we naturally get excited. We want to implement it everywhere. And yet, the new intervention so often fails in new settings. It isn’t as effective as advertised and/or it’s more expensive. The intervention simply does not scale-up and potentially results in harm. Effort and resources are diverted from those things that already work better there to implement the new intervention, which showed so much promise in the original setting, here.

The intervention does not fail in new settings because the cause-effect never existed. It fails in new settings because causes are local. The effect that was observed here was not caused by A alone. The intervention was not a singular cause. A causes B within a context that allows the relationship between cause and effect to be manifest. The original research in which A was identified had social, economic, cultural, political, environmental, and physical properties. Some of those properties are required for the realisation of the cause-effect. This means that generalisation is really about re-engineeering context. We need to make sure the target settings have the the right contextual factors in place for the intervention to work. We are re-creating local contexts. The implementation problem is one of understanding the re-engineering that is required.