Global South risks losing out as Big Tech crowds out local AI tools

A few days ago, a message made the rounds that Perplexity AI has launched a fact-checking assistant on WhatsApp. To many, it was just another announcement in the fast-moving world of artificial intelligence, but to those working in the information ecosystem, particularly in the Global South, it felt like something more: a sign of just how quickly Big Tech is expanding into domain-specific spaces once led by local innovation.

We are entering a new phase of AI development, one where tools are no longer just general-purpose assistants but are being fine-tuned to sectors like health, education, agriculture, journalism, and law. This sectoral shift holds immense promise. But it also revives an old pattern: the consolidation of power in the hands of those who control infrastructure, not just ideas.

Across the Global South, local innovators have built thoughtful, culturally aware AI systems tailored to real needs. Fact-checking tools that understand local slang. Legal chatbots that explain tenancy laws in plain English. Health apps that adjust for infrastructure gaps. These aren’t just tools—they are acts of epistemic justice. They reflect an understanding that intelligence is not objective or universal. It is shaped by language, history, geography, and power.

But just as these efforts take shape, they are being outpaced, if not overshadowed, by Big Tech’s arrival.

Stay Ahead with Premium Times

Add as a preferred source on Google

Global companies are now deploying sector-specific AI tools deeply embedded in dominant platforms. Perplexity’s assistant lives directly in WhatsApp, one of the most widely used apps in Africa. Google’s Med-PaLM is being trialed in hospitals, aiming to become the standard for AI-assisted medical decision-making. Khan Academy’s Khanmigo, built on GPT, is pitched as a global tutor. The question is not whether these tools work; it is about who decides what counts as knowledge, who owns the interface to the citizen, and whose values are silently embedded in these systems.

Even in areas where we have argued for agency—where critical knowledge work, like journalism, writing, and research, are not victims of AI but sit at its power core—there is a need to pause and reflect. I have previously made the case that media actors must not retreat in fear but embrace AI as a lever of influence. But influence does not happen in a vacuum. Without structural equity and algorithmic fairness, even the most contextually grounded local tools risk becoming invisible, overshadowed not by merit but by scale. In other words, we can lead in thought and design, yet still lose ground if the platforms and rules are rigged.

This moment demands a sharper understanding of digital sovereignty—not as a nationalist impulse, but as a form of collective agency. It is the ability of people, communities, and countries to shape the digital tools they use, to ensure those tools reflect their histories, languages, needs, and aspirations.

Without it, we risk repeating old patterns in new clothes. In the colonial era, infrastructure was used to extract value while marginalising local systems. Today, if we are not careful, AI could become the railway and telegraph of the digital age—extracting data from the South, building intelligence in the North, and redistributing “solutions” that don’t always fit our realities.

Without digital sovereignty, AI becomes a new form of dependency, just as colonial railways and telegraph lines moved people and goods for the benefit of imperial centres, today’s AI infrastructures—APIs, platforms, and datasets—risk extracting our data, flattening our differences, and returning “intelligence” built elsewhere, for elsewhere.

Already, we see how local civic tech tools, many of them funded through grants or developed by journalists and civil society actors, struggle to compete once global platforms roll out similar features. Their visibility drops. Donor interest wanes. Governments are told to “plug in” to global APIs. Over time, this weakens the local ecosystem. We lose not just tools, but the capacity to build, adapt, and govern.

The implications are profound when health systems rely on AI trained on foreign data, when legal guidance comes from models trained in different legal cultures, and when students learn from tutors who don’t understand their world. We are not just using tech, we are surrendering control.

This isn’t abstract; it is already happening as the most contextually rich, low-resource language models are struggling to find investment. Civic tech tools developed by African organisations are often seen as duplicative once a global model releases a generic version.

Epistemological Danger

When knowledge is generated by tools trained largely on data from the Global North, there is a quiet reordering of authority. Who gets to define what is true? Whose definitions of fairness, credibility, or safety are baked into the system? When the answers are decided elsewhere, we’re not just consuming foreign knowledge—we are outsourcing judgment.

This dynamic is not unique to AI. We saw it with social media, where algorithms shaped virality, discourse, and even elections. But with AI, the stakes are deeper. AI doesn’t just distribute content; it generates answers, predicts outcomes, and defines what is “reasonable,” “plausible,” or “appropriate.” And increasingly, it does so invisibly in high-stakes areas like health, education, and governance.

This is not a call to resist all global innovation as digital sovereignty is not isolationism; it is about being at the table—shaping, negotiating, and regulating from a place of informed strength.

What we need now is a multi-layered response:

Investment in public-interest AI, including low-resource language models, locally built datasets, and civic-use cases that reflect diverse realities.

Policy and procurement frameworks that mandate ethical use of AI and prioritise local solutions where relevant.

Open infrastructure and shared compute: so civic innovators and public institutions are not locked out of the AI economy.

South-South collaboration to pool expertise, share standards, and co-develop tools aligned with regional values.

READ ALSO: African Newsrooms in Age of AI: Forging Strategic Partnerships for Compensation

This also calls for new narratives, ones that resist the seductive idea that intelligence only comes from scale and instead affirm that context, care, and consent are equally powerful inputs in building meaningful AI.

Just as algorithms shaped virality during the social media boom, rewarding outrage, sensationalism, and speed, AI is now shaping what becomes truth, what gets amplified, and who gets heard and if we do not interrogate the power structures behind AI deployment, we will wake up to a world where our tools reflect someone else’s values, our facts are filtered through someone else’s lens, and our innovators are edged out before they can even begin.

This is not just a tech issue; it is a question of epistemic power, democratic accountability, and the digital futures we want to build.

We still have time to act, but only if we recognise that the future of AI is not just about automation or scale; it is about whose intelligence counts.