Training AI with Silence: What the Global South Can Teach the Algorithm
In most of the world, artificial intelligence is taught to listen to noise. It consumes millions of voices, scrolls endlessly through conversations, headlines, reactions, and ratings. By Tuhu Nugraha
The Western approach to AI training focuses on acceleration – more data, faster computation, and sharper models. However, in the Global South, a different approach is emerging. One that is slower and quieter.
The Kind of Data That Doesn’t Trend
In Indonesia, stories are told about people who don’t go viral. A satpam (security guard) watches as everyone walks past him while he guards a smart office he can’t afford to enter. A mother in Indomaret, a ubiquitous chain of neighborhood convenience stores across Indonesia, stares at a digital promotion for a product she cannot pronounce. These stories offer a radical alternative to the typical AI training data: texture, ambiguity, and humanity.
By personifying objects like plates, security gates, or display screens, critique becomes softer and more powerful. This approach reduces defensiveness and invites reflection rather than resistance. In the Global South, where critique can easily trigger backlash, metaphor becomes a safer way to advocate for change.
What Silence Trains Better Than Speed
When AI is trained solely on statistical accuracy and optimized feedback loops, it learns to answer fast. But it doesn’t learn to pause, to listen deeper, or to question its responses. Narrative, especially from the Global South, offers contradictions and carries memory, shame, resilience, and irony. It’s not about teaching AI facts, but about teaching it to carry a question without rushing to resolve it.
The narrative series ‘P.S. from the South’ writes letters from the Global South, speaking in objects, silence, and memory. It’s a space for things that feel but rarely get said. This approach is a path of dialogue that doesn’t shout, doesn’t seek dominance, and doesn’t provoke through force. It stirs change through dignity and reflection, not violence or spectacle.
Why the South Must Influence the Algorithm’s Voice
We are no longer just users of AI; we are its subjects, consumers, mirrors, and datasets. If the voice of the Global South doesn’t shape the algorithm, the world will see us only through the eyes of those who have never truly heard us. AI is now a lens through which future generations will be understood. If we remain absent from that lens, we will be misrepresented or rendered invisible.
What the South Teaches the Algorithm
The South teaches AI not to assume and not to automate empathy. It teaches that not all patterns deserve prediction and that context matters. Some truths are expressed not in words but in tone, in the decision to remain quiet, or in the absurdity of certain situations.
Narrative as Resistance, Prompting as Diplomacy
In a time where data is currency, storytelling becomes subversive. Teaching a machine using the voice of the marginalized, the overlooked, and the poetic is an act of narrative diplomacy. It says, ‘This is also intelligence. This is also real. This is also worthy of being known.’ Through stories, we prompt the algorithm to reflect. Through silence, we invite it to grow.
Conclusion: Toward a More Reflective AI
The Global South is training AI not with machines, but with meaning. Not with speed, but with stillness. Not with dominance, but with dignity. Perhaps what we need most is Artificial Generous Listening – the kind that starts with sitting beside a story and asking nothing except to be allowed to feel it fully. And that might be what truly teaches a machine to be intelligent.
