This is Part 5, the final piece in a five-part series on artificial intelligence, seen through Rwanda's experience of building its institutions from scratch. Last week's piece looked at AI competing for our affection, not just our attention. This week closes the series with a question about something even older than technology: responsibility. Solange read the post twice before sharing it. It was well written, confident, exactly the kind of thing her friends would agree with. She never asked who wrote it. She couldn't have found out if she tried. It could have been a neighbor down the road. It could have been a programme that had never set foot in Rwanda or met a single reader. That's the strange thing about this new wave of AI. It doesn't need a visa or a plane ticket to reach you. It doesn't cross a border checkpoint. It just appears, already fluent, already capable of writing convincingly about places and people it has never known. The historian Yuval Noah Harari calls this an immigration wave with no visas at all. This one doesn't move into our neighborhoods. It moves into our conversations, our institutions, and eventually our heads. The pamphlet that outran its own author Go back to 1517, to a German monk named Martin Luther. He nailed a list of complaints to a church door in Wittenberg, arguing that the Church had lost its way. He expected an argument with local clergy. Instead, a new invention, the printing press, carried his words faster and farther than any single person could. Within weeks, Luther's arguments were being read in cities he had never visited. Luther's words could travel without Luther. But the voice behind them was still recognisably his. A reader could ask who wrote them and arrive at the same monk a bishop could summon and question. An idea has never needed a passport to cross a border. What's new is how difficult it can be to find someone prepared to answer for it. The voice no one answers for AI can already produce fluent Kinyarwanda without ever having lived a single Rwandan experience. It can write about family without having parents, faith without having prayed, grief without having buried anyone, and Rwanda without ever calling it home. The problem isn't that nobody is behind those words. Often, many hands are involved — the people who built the model, the material it learned from, whoever prompted it, whoever edited it, and whoever finally published it. AI doesn't eliminate authorship. It fragments it. That is precisely the kind of problem institutions are supposed to solve. Rwanda has spent thirty years rebuilding around a simple expectation: responsibility should eventually have an address. A decision has an official behind it. A judgment has a court. A newspaper has an editor. AI makes that address harder to find. A statement can begin in training data, pass through a model, be shaped by a prompt, selected by one person, edited by another, and reach thousands before anyone asks who will answer for it. But responsibility doesn't end with the person who produced the words. It also belongs to those of us who receive them. Human beings have always lived among other people's ideas. We inherit language from parents, beliefs from communities, arguments from books, teachers, and friends. The danger was never that AI puts someone else's thoughts in your head. It's that it makes it harder to notice when you've stopped thinking and started agreeing. The answer isn't to trust every name attached to a statement. People with names can lie, and unsigned words can be true. It is to ask where a claim came from, what supports it, and who is willing to stand behind it when challenged. And before we pass it on, we must ask whether we are willing to do the same. Solange never found out who wrote the post she shared. Perhaps she never will. AI can scatter authorship across programmers, models, prompts, and people she will never meet. But sharing was her decision. Whatever the origin of those words, passing them on made her part of their journey. Responsibility had found another address. Hers. The writer is a Professional Engineer specializing in process design. This concludes his five-part series on artificial intelligence and Rwanda's institutions, from a driver who ignored an app to a monk whose words outran him.
