Welcome to Issue #74 of TAIS, where every Friday we spotlight visionary changemakers reshaping Africa’s tech, data, and AI landscape, one breakthrough at a time.
Today we spotlight Dr. Adio-Adet Dinika, a Zimbabwean political scientist and researcher at ZeMKI, University of Bremen, and the Distributed AI Research Institute, whose path into AI labour scholarship began not in a seminar room but on Fiverr and Upwork, competing for work in a market that promised freedom and delivered risk. His world operates where platform labour meets political economy, where an unpaid hour spent pitching, an opaque rating, or a sudden rule change determines whether a worker stays visible or disappears entirely and where the people annotating, moderating, and correcting the data behind every AI system are treated as scaffolding to be removed rather than the knowers who make the system function. Not in the space where AI is celebrated for what it can do, but in the harder, less visible layer beneath it: who mined the minerals, who labelled the data, who absorbed the traumatic content, and who was never given authority over any of it.
Dr. Dinika is not building models, evaluating them, or governing them from the outside. He is arguing that the real story of AI is the labour hidden underneath it, and that hiding it is not accidental but commercially useful, a system looks more autonomous, and more investable, once the human work behind it disappears. His argument, which runs through every answer in this conversation, is that recognition without authority is just another form of extraction: workers can be featured in a report or thanked in a model card while holding no power over wages, categories, or redress. He refuses to treat AI as neutral infrastructure, insisting it is “society sedimented into an object” and that calling it inevitable is itself a political move, one that turns contestable choices into natural events.
His work through the Data Workers’ Inquiry and the Possible Futures series is his answer to a question the industry rarely pauses to ask: not just where the human is in AI, but what power that human actually holds and whether the workers closest to the harm will ever be given authority over the systems they sustain.
Q: You’ve written that your interest in digital labour emerged not only from theory, but from your own experience working on platforms like Fiverr and Upwork after graduate school. How did moving through those systems personally reshape the way you understood technology, labour, and power?
A: My interest in platform labour began as biography before it became research. After graduate school, I turned to platforms such as Fiverr and Upwork to earn an income, carrying the familiar promise: that digital platforms remove borders, reward talent and let anyone with an internet connection into a global market.
What I found was more complicated. I was competing with thousands of workers across radically unequal economies, often underbidding simply to become visible. A client could vanish without explanation, a single rating could decide whether the next client ever saw me, and the platform could change the rules while presenting itself as a neutral intermediary. Much of the work was unpaid, searching, pitching, revising profiles, waiting, managing my reputation. What was advertised as freedom mostly meant that the worker carried the risk.
That experience changed how I understood technology. I stopped seeing platforms as tools connecting buyers and sellers and began seeing them as private systems of governance: they decide who becomes visible, what counts as good performance, how disputes are resolved and whose account is suspended. My so-called villain origin story was realising that the market had not disappeared into code , power had. That recognition shaped my doctoral research on platform work in Rwanda, South Africa and Zimbabwe, and it still informs how I study AI labour today.
Q: You argue that technology should not be understood as neutral infrastructure, but as a site where power, capital, and history converge. What do mainstream AI conversations still fail to recognise when they frame these systems as neutral or inevitable?
A: Mainstream conversations often begin too late. They start with a model that already exists and ask whether it is accurate, safe or biased. The more fundamental question is how the model became possible: who financed it, whose data trained it, which workers cleaned and interpreted that data, what infrastructures sustain it, and who benefits from its deployment.
Technology is never simply an object placed into society; it is society sedimented into an object. Design choices carry assumptions about intelligence, normality, efficiency and value. Datasets carry histories of inclusion and exclusion. Business models shape what problems get investment, and procurement decisions determine which technologies become public infrastructure. Even the language of “scale” tends to reflect the priorities of capital rather than the needs of communities.
The claim that AI is inevitable is especially dangerous, because inevitability turns political decisions into natural events. It tells workers to adapt, governments to compete and communities to accept whatever disruption arrives. But AI systems are built through choices , about ownership, labour, energy, data, deployment and accountability, and so they can be contested and changed.
When people call technology neutral, they usually mean the politics have been hidden successfully. My work tries to put those politics back into view. The question is not only what an AI system can do, but who decides what it should do, who bears its costs, and who acquires power once it is embedded in everyday life.
Editorial Commentary: Dr. Dinika’s experience points to a less obvious shift in how digital work distributes risk. Platforms can make a labour market look more open while quietly moving more of the uncertainty onto the worker: finding work, proving credibility, absorbing downtime, and adapting when the rules change. The promise of flexibility can therefore coexist with greater insecurity, because the platform gets to organise the market without necessarily sharing the risks that come with it. That distinction matters as digital labour expands beyond freelancing into the wider AI economy. The more work is mediated by platforms, algorithms, and automated systems, the easier it becomes to describe these arrangements as efficient or flexible while overlooking who absorbs the instability underneath. What changes with the platform is not necessarily the insecurity itself, but where that insecurity sits. The worker still carries the uncertainty of finding work, maintaining a reputation, and adapting to changing rules; the difference is that much of this is now mediated by a system that presents those conditions as simply how the market works. That becomes particularly important as AI begins reorganising more kinds of work.
Q: Your work examines how AI infrastructures can reproduce older patterns of extraction under the language of innovation. When you look at today’s AI economy, what continuities do you see between digital systems and earlier imperial or colonial structures?
A: I would not claim that every digital relationship is identical to colonial rule, history deserves more precision than that. But there are unmistakable continuities in the structure of extraction.
Colonial economies organised territories as sources of raw materials and labour while concentrating ownership, processing, knowledge and profit elsewhere. The AI economy can reproduce that geography: minerals extracted in one place, data collected in another, content annotated and moderated somewhere else, while models, intellectual property and profits concentrate in a handful of corporations and countries. African societies are frequently positioned as sources of data, labour, minerals and expanding markets rather than as sovereign producers and governors of technology.
There is also an epistemic continuity. Colonial systems did not extract only material resources; they classified people and imposed the categories through which the world was governed. AI depends on classification too. The people closest to a context may see that a category is culturally wrong, politically dangerous or ethically inadequate, yet they are rarely given authority to change the taxonomy. Knowledge travels upward while decision-making stays elsewhere.
The language has changed , extraction now arrives dressed as innovation, inclusion or development , but inclusion into an unequal value chain is not sovereignty over it. So the decolonial question is not whether Africa is present in AI. Africa is already deeply present. The question is whether African workers, communities and institutions hold meaningful power over the infrastructures, knowledge and value they help create.
Q: Public discussions about AI often focus on models and automation, while the labour systems behind them remain largely invisible. What kinds of human labour currently sustain AI systems, and why do you think that labour is so consistently obscured?
A: AI is sustained by an enormous range of human labour: workers who mine the minerals in our devices and data infrastructure; people who build and maintain data centres, cables and energy systems; engineers and researchers; workers who collect, transcribe, translate and clean data; annotators who label images, speech and text; evaluators who compare model outputs; content moderators who process violent and disturbing material; and users whose interactions continuously generate feedback.
Much of this is described as low-skilled, yet it demands sophisticated judgement. A worker deciding whether a phrase is hate speech must read language, culture, context, irony and power. Someone annotating an image may have to interpret practices or objects the client’s taxonomy barely recognises. Workers are not supplying hands to AI; they are supplying interpretation.
This labour is obscured partly because the mythology of automation is commercially useful: a system looks more advanced, autonomous and investable once the human work behind it disappears. Long subcontracting chains put distance between major technology companies and conditions on the ground, and non-disclosure agreements, fragmented contracts and platform interfaces make workers hard to see and harder to organise.
The invisibility is not accidental; it performs economic and ideological work, keeping labour costs low while letting companies market accumulated human intelligence as machine intelligence. The scandal is not only that workers are underpaid. It is that their knowledge is absorbed into systems that then deny their role as knowers.
Editorial Commentary : Dr. Dinika's most interesting point may be what happens to the meaning of "AI" once we take its human labour seriously. We tend to imagine intelligence entering the system through the model, while the people who teach it how to interpret language, images, behaviour, and context are treated as if they are simply preparing the raw material. But the examples he gives complicate that distinction. When a worker decides whether a phrase is hateful, recognises irony, or interprets something through a particular cultural context, they are not merely labelling data; they are exercising judgement that the system subsequently learns to reproduce. Some of what we describe as machine intelligence is therefore accumulated human interpretation that has been separated from the humans who supplied it. That changes how we should think about the workers behind AI. Their significance is not exhausted by the wages they receive or the conditions under which they work. They are participating in the production of what the system will eventually present as its own capacity to understand. And once that distinction disappears, the question of who built the intelligence becomes much harder to separate from the question of who gets recognised as having intelligence in the first place.
Q: You’ve described platform work as a labour market shaped by opacity, algorithmic ranking, and relentless underbidding. How do platform systems restructure workers’ relationships to visibility, competition, and survival?
A: In a conventional labour market, you may know the employer, the wage structure and at least some of the rules of evaluation. Platforms replace much of this with an interface and an algorithm. Workers know visibility matters but rarely how it is produced: ratings, response times, acceptance rates, prices, past earnings and undisclosed signals can decide who surfaces in a search and who effectively disappears.
Visibility thus becomes a scarce resource the platform allocates, and workers are pushed to treat one another as competitors in a permanent global auction. The easiest lever is usually price, which drives relentless underbidding, yet a lower price can also signal lower quality, while refusing poorly paid work reduces activity and future visibility. The worker is trapped inside rules that are both powerful and hard to inspect.
Survival becomes reputational. A bad rating is not feedback on one transaction; it can damage access to all future work. Fear of losing visibility pushes workers to accept extra revisions, tolerate abusive clients or stay constantly available. The language of entrepreneurship masks a relationship in which workers must continuously manage themselves to satisfy opaque platform demands.
This is why platform autonomy is contradictory. You may choose when to log in, yet have little control over prices, allocation, evaluation or account security. The platform does not remove management; it makes management less visible, more automated and harder to contest.
Q: You’ve argued that every AI system depends on global supply chains stretching across unequal geographies. What becomes visible when we stop thinking about AI as “software” alone and instead examine the full chain of labour and infrastructure behind it?
A: The first thing that becomes visible is materiality. AI does not descend from the cloud. It runs on mines, chips, warehouses, data centres, electricity grids, water, undersea cables, logistics networks and human bodies. Follow that chain and the clean image of weightless software gives way to a political economy of land, labour, energy and extraction.
We also see how costs and benefits are geographically split. A community may bear the environmental cost of mineral extraction or a data centre’s thirst for water and power without sharing in the value it generates. A worker may perform the culturally complex task of interpreting data while intellectual property and profits accumulate elsewhere. When labour resistance or regulation rises in one jurisdiction, companies shift contracts to another: the supply chain gives capital mobility, while workers stay embedded in particular legal and economic conditions.
Following the full chain also widens what counts as AI governance. It cannot be confined to model performance or abstract ethical principles; it has to include labour rights, environmental impact, public infrastructure, taxation, procurement, data ownership, community consent and corporate responsibility across subcontracting chains.
The phrase “AI system” tempts us to picture a bounded technical object. In reality AI is a relationship among institutions, workers, communities, resources and environments , and once that relationship is visible, accountability can no longer stop at the company whose logo is on the final product.
Q: Your research includes fieldwork in South Africa, Rwanda, and Zimbabwe. What specific dynamics shape digital labour economies in African contexts, particularly within global AI supply chains?
A: The first point is that there is no single African digital labour market. Different political economies shape how workers meet platforms: Rwanda’s coordinated digital-development agenda, South Africa’s comparatively developed technological and industrial base, and Zimbabwe’s long experience of volatility and informality create distinct conditions. Still, several dynamics recur.
High unemployment and underemployment make platform work attractive, especially to young people seeking a way into global markets. Weak social protection then means workers absorb the costs of equipment, connectivity, illness, fluctuating demand and unpaid time, with internet and device costs a real barrier. Currency instability can make foreign-denominated earnings valuable, but payment restrictions and exchange-rate swings can just as easily make income unpredictable.
Global platforms arrive inside these inequalities. Workers compete across borders without equal infrastructure, legal protection or bargaining power. Linguistic and cultural knowledge can be highly valuable to AI supply chains while remaining poorly paid, and rural workers, women and people with limited connectivity face added exclusion.
African states also face a development dilemma. Digital labour is promoted as a route to employment and global participation, while the quality of those jobs and the distribution of value get far less attention. The challenge is not simply to attract outsourced work. It is to secure decent work, build local technological capacity, and keep African economies from being permanently assigned the lowest-paid sections of the AI value chain.
Editorial Commentary: Dr. Dinika complicates what we mean when we say Africa is participating in the AI economy. Being part of the supply chain does not tell us where the value sits within it. A country can attract AI-related jobs, provide data or minerals, host infrastructure, and still have little influence over the platforms, intellectual property, standards, or profits that sit further up the chain. His comparison across South Africa, Rwanda, and Zimbabwe also matters because it shows that there is no single African route into this economy; national conditions shape what workers can actually gain from participation. That distinction is easy to lose when AI investment itself is treated as the measure of progress. More contracts, more outsourced work, or more infrastructure may signal integration without necessarily building the capabilities that allow economies to negotiate better terms later. What matters is where that participation leaves African economies. If it builds local capabilities, ownership, and bargaining power, integration into the AI economy can become a foundation for something bigger. If it does not, Africa may simply become more deeply embedded in systems whose most valuable assets remain elsewhere.
Q: Many current AI governance discussions focus on regulation, safety, or technical standards. From your perspective, where do governance frameworks still fall short in addressing questions of labour and economic power?
A: Many governance frameworks are model-centric. They ask whether a system is accurate, explainable, secure or discriminatory , all important , but rarely whether it was produced through exploitation or whether its deployment concentrates economic power.
Labour appears, if at all, as a secondary social impact. Data workers, content moderators and platform workers are seldom treated as governance actors, yet they meet the problems first: defective categories, recurring errors, traumatic material and cultural distortions surface during production. Current frameworks harvest this knowledge as quality control but rarely give workers authority to pause a project, demand a redesign or enter their concerns into the formal model record.
Governance therefore has to move upstream and outward. It should carry enforceable standards for wages, psychological support, collective bargaining, freedom of association and the right to refuse dangerous work. Lead companies should answer for conditions across their subcontracting chains instead of outsourcing both labour and liability, and public procurement should require disclosure of labour practices, not only technical performance.
We also have to confront ownership and concentration. A technically safe system can still deepen dependency if a few firms control compute, data, platforms and standards. Regulation that manages harms without redistributing power risks making an extractive system tidier rather than more just. Governance should not only ask how to control AI; it must ask how control over AI is distributed.
Q: Workers involved in annotation, moderation, data labeling, and platform labour are often treated as peripheral to AI innovation. How should we rethink the role of these workers within the broader AI ecosystem?
A: We should start by dropping the idea that these workers merely prepare raw material for the “real” work of innovation. Annotation, moderation and evaluation are forms of knowledge production: workers interpret ambiguity, apply cultural judgement, catch errors and translate complex realities into categories a machine can process. Without this labour, many systems we call intelligent would not function.
The shift I advocate is from labour as input to labour as authority. Fair wages, secure contracts and mental-health protections are essential, but recognition cannot stop there. Workers should have channels to challenge harmful instructions, contest inadequate taxonomies, document recurring problems, refuse dangerous tasks and take part in decisions about how systems are built and deployed. Their knowledge should enter audits, model documentation and risk assessments.
This is the principle behind the Data Workers’ Inquiry: those who do the work should not merely be research subjects. They should be able to investigate their own workplaces, set the questions that matter and speak with epistemic authority. Too often institutions consult workers only after the key decisions are made. Genuine participation means power before and during production, not after.
Workers are currently close enough to produce intelligence but too far to claim authority over it. A just AI ecosystem would correct that contradiction, recognising them not as peripheral service providers but as co-producers whose labour, knowledge and rights are foundational to the technology.
Q: Terms like “innovation,” “flexibility,” and “future of work” are frequently used to describe AI-driven economies. What realities can those narratives obscure?
A: These terms are powerful because they sound unquestionably good. Who wants to oppose innovation or flexibility? Yet they often work as containers into which very unequal arrangements are quietly placed.
“Flexibility” can mean a worker chooses when to work , or that the company guarantees no hours, income, insurance or long-term responsibility. “Entrepreneurship” can describe genuine independence, or rename a worker who carries business risk without real control over prices or conditions. “Innovation” can celebrate a new product while erasing the labour practices, mineral extraction and public subsidies that made it possible.
The “future of work” narrative is especially evasive. It presents technological change as something arriving from outside society, to which workers must adapt , obscuring that firms and governments actively choose which technologies to fund, where to deploy them and how to share the gains. The future is not an autonomous force; it is negotiated, financed and governed.
These narratives also count jobs created while ignoring their quality. A new digital occupation may still bring poverty wages, unpaid waiting, psychological harm, surveillance and no social protection. A job can be new and still be old exploitation in digital clothing.
The task is not to reject innovation but to make the word answerable: innovation for whom, under whose control, at whose expense, and toward what kind of future?
Editorial Commentary: There is a strange hierarchy built into the way AI governance currently understands expertise. A worker can be trusted to identify a harmful category, flag a recurring error, or make a difficult judgement about context, yet that same worker is rarely treated as someone with authority over the system producing those problems. Their knowledge becomes evidence for an audit rather than a reason to give them a seat in the decision itself. Dr. Dinika's argument exposes the contradiction: the closer you are to the practical reality of an AI system, the more useful your knowledge may be and yet that does not necessarily give you more power over it. If governance continues to treat these workers primarily as people to protect from AI rather than people whose knowledge should shape AI, it risks reproducing the same hierarchy it claims to regulate.
Q: Through your work with the Distributed AI Research Institute and the Possible Futures series, you engage with questions of alternative technological futures. What would a genuinely just AI future require beyond technical improvements alone?
A: A just AI future cannot be reached by dropping a slightly fairer model into the same extractive political economy. Technical improvements matter, but justice requires changing the relationships around the technology.
That means redistributing authority to workers and communities, not merely consulting them: public and cooperative alternatives to concentrated ownership; real control over data and language resources; access to compute and technical capacity; strong labour protections; environmental limits; and institutions capable of democratic oversight. African societies should not have to choose between technological participation and technological sovereignty.
Through Possible Futures I work with three ideas , repair, refusal and reworlding. Repair means addressing the material and epistemic debts created by extraction. Refusal preserves the right of workers and communities to reject harmful categories, deployments and forms of capture. Reworlding asks us to build technologies organised around reciprocity, care, public value and collective flourishing.
The futures I find compelling are not distant utopias but politically achievable tomorrows: data workers with governance rights over the systems they train; African-language technologies in which communities control the data and share the benefits; public AI infrastructures built by universities, cooperatives and public-interest institutions; systems designed with affected people as political subjects rather than data points.
Big technology companies have colonised not only infrastructure but imagination, making their dominance look inevitable. A just future begins by recovering our capacity to imagine , and to institutionally build , otherwise.
Q: You position yourself as a decolonial scholar of AI labour and governance. What does a decolonial approach make visible about AI systems that more conventional approaches often overlook?
A: A decolonial approach changes the unit of analysis. Instead of treating a harmful output as an isolated technical error, it asks about the historical and global relations that produced the system. Who defines intelligence? Whose knowledge becomes data? Who designs the categories? Where does value travel? Which societies are expected to supply labour, minerals and data, and which keep ownership and authority?
It also makes epistemic power visible. AI systems do not merely process the world; they classify it, and those classifications can universalise assumptions formed in particular cultural and institutional contexts. A worker or community may see that a category cannot describe their reality, yet conventional governance treats that knowledge as anecdotal rather than authoritative.
Decolonial work is therefore not a request for symbolic representation, nor the addition of African examples to an unchanged framework, nor is it anti-technology. It is a demand to transform the relations through which technology is imagined, produced and governed , to recognise Africa not simply as a source of data or a site of deployment, but as a producer of theory, technological vision and institutional alternatives.
For me this connects directly to digital Pan-Africanism: the effort to join collective technological self-determination with democratic rights and accountability. Sovereignty without rights can arm states against citizens; rights without material capacity remain declarations without force. A decolonial approach insists on holding together ownership, justice, history and the freedom of African societies to shape their own technological futures.
Q: As AI systems become increasingly embedded into economic and institutional life, what developments concern you most, and where do you still see room for intervention or transformation?
A: I am most concerned by the concentration of infrastructure and decision-making power in the hands of the broligarchy — a small circle of firms, and the men who front them, that now controls the compute, cloud services, models, platforms and standards on which public and private institutions depend. This is an infrastructural power ordinary consumer choice cannot touch, and it is increasingly personalised. We are invited to admire a handful of AI superstars; the founders, the celebrity researchers, the faces on the magazine covers, as though these systems were feats of singular genius rather than the product of millions of hands.
I am equally concerned by AI’s spread into high-stakes domains, for employment, welfare, migration, policing, education, public administration, where a single classification can decide whether someone gets an opportunity, becomes a suspect or is denied a right. Nowhere is this starker than at the border. Systems that score “risk,” match biometric data and triage asylum claims increasingly decide who is waved through and who is turned away and those turned away are usually the people with the least power to contest the decision. An asylum seeker fleeing real danger can be reduced to a probability and refused before a single human being ever hears their story. These systems make political judgements look administrative and can scale one error across entire populations. And the applause for the superstars depends on a silence: the annotators, moderators and data workers in Nairobi, Gulu, Accra and across the global majority who clean the data, absorb the traumatic content and carry the psychological and ecological costs, while the profit, the prestige and the story accumulate elsewhere. The broligarchy is celebrated for an intelligence that was, in large part, produced by the very workers it refuses to name.
But I do not believe the future is closed. Workers are organising and documenting their conditions. Civil society is building governance models from below. Public institutions can use procurement to demand labour transparency and community accountability. African countries can cooperate regionally on infrastructure, standards and bargaining power rather than negotiating one by one with global firms. Universities, cooperatives and public-interest organisations can build alternatives around local languages and public value.
My work with the Data Workers’ Inquiry and AI Pan-Africanism rests on this possibility. Those closest to harm are often closest to diagnosis, and transformation becomes possible when their knowledge is turned into institutional power: the power to pause, contest, redesign and, where necessary, refuse.
Q: What is something about AI labour, governance, or platform economies that you think remains widely misunderstood?
A: The most persistent misunderstanding is that exploitation is a temporary defect of an otherwise autonomous technology. We are told AI systems are becoming more independent and that the remaining human labour will gradually disappear, as if workers were scaffolding to be removed once the machine is finished. In reality, automation usually reorganises human labour rather than eliminating it. New systems create new tasks of data production, evaluation, moderation, correction, maintenance and supervision. Some of that work becomes more visible; much of it is pushed into subcontracting chains, platforms or unpaid user activity. The disappearance is often representational rather than material: the worker remains, but the story stops including them.
A related misunderstanding is that recognition just means making hidden workers visible. Visibility matters, but it is not enough. A worker can be featured in a report, thanked in a model card or invited to a consultation while holding no power over wages, categories, deployment or redress. Recognition without authority becomes another form of extraction.
So the central question is not only “Where is the human in AI?” but “What power does the human hold?” We need to move from labour as input to labour as authority. Until workers and affected communities can shape the systems they sustain, AI governance will keep listening to the people closest to harm only after the most important decisions have already been made.
Editorial Commentary: What makes Dr. Dinika’s idea of a just AI future interesting is that he is not asking us to make the existing system slightly less harmful. He is questioning whether the system we have accepted as inevitable should be the starting point at all. His language of repair, refusal and reworlding moves the conversation from better regulation toward the harder work of building different arrangements of ownership, authority, infrastructure and value. That matters because so much of AI governance begins with the systems already in place and asks how to make them safer, fairer, or more accountable. Dr. Dinika asks what becomes possible when we stop treating those arrangements as fixed. For Africa, that distinction is particularly important. Technological sovereignty cannot mean simply owning more infrastructure if the underlying relationships remain extractive; it has to include the capacity to decide what gets built, whose knowledge counts, and what technology is ultimately for.
Dr. Dinika brings a different starting point to the AI conversation. We tend to encounter it at the point where a model produces an answer, a platform makes a decision, or an automated system enters an institution. He keeps taking us further back, to the workers whose judgement becomes part of the system, the infrastructures that make it possible, the histories that shape its categories, and the institutions that decide who gets to exercise power over it.
That perspective also changes what a just AI future would mean. It cannot be reduced to better models, more representative datasets, or stronger safeguards if the underlying distribution of ownership and authority remains unchanged. For Dr. Dinika, the harder work is building the capacity to decide, contest, refuse and create alternatives. His contribution to this conversation is therefore not simply a critique of the AI economy, but an insistence that the people who sustain these systems should have a meaningful role in determining what they become.
And perhaps that is the most important question he leaves us with: not whether humans will remain somewhere inside AI, but whether they will have enough power to shape the systems they are helping to build.
Thank you for reading!
Don’t see your pick in the options? Drop it in the comments. Dr. Dinika joins the map this weekend.
Thank you for standing with this work.
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.