--- title: "Prompting Alone: What We Lose When We Stop Asking Each Other for Help" authors: ["Hafiz Hanif"] date: 2026-08-15 doi: "10.5281/zenodo.21941765" license: "CC-BY-4.0" language: "en" canonical: "https://hackedu.tech/articles/prompting-alone-what-we-lose-when-we-stop-asking-each-other-for-help" keywords: ["Technology", "participation", "help", "Artificial Intelligence", "prosocial behaviour"] --- # Prompting Alone: What We Lose When We Stop Asking Each Other for Help > **Cite as:** Hanif, H. (2026, August 15). Prompting Alone: What We Lose When We Stop Asking Each Other for Help. Hack(edu).tech Chronicle. https://doi.org/10.5281/zenodo.21941765 > **DOI:** https://doi.org/10.5281/zenodo.21941765 ## Abstract Online help communities are experiencing measurable decline as users migrate from human forums to generative AI. Drawing on Robert Putnam's social capital thesis and the author's doctoral research on helping behaviour in open online communities, this article examines the trade underway: AI offers instant, judgement-free assistance, removing the social conditions that human help has always carried. Yet those same conditions built membership, reputation, and shared knowledge commons. Empirical evidence shows information-exchange platforms collapsing while socially-bonded communities endure. The article argues this shift is irreversible but its terms deserve scrutiny, particularly as solitary, AI-mediated learning enters classrooms. --- In 2000, the political scientist Robert Putnam published Bowling Alone, a book that documented the quiet collapse of American community life. Americans were still bowling, he observed, but no longer in leagues. The activity survived; the togetherness did not [@putnam2000]. A quarter of a century later, I find myself watching the same story unfold in a place Putnam never studied: the online forum. For most of the Internet's life, when you were stuck on a problem, you asked people. You posted your half-broken code on Stack Overflow, your obscure error message on a bulletin board, your beginner's question on a community forum, and you waited. Strangers, owing you nothing, showed up and helped. This was one of the quiet miracles of the connected age, and it is disappearing in front of us. The evidence is no longer anecdotal. Within six months of ChatGPT's release, activity on Stack Overflow fell by an estimated 25 percent relative to comparable platforms where access to the tool was limited, and the decline was sharpest in exactly the areas where the AI performs best [@delriochanona2024]. A separate study found significant drops in both visits and question volumes on Stack Overflow, with average answer quality declining as experienced contributors quietly exited [@burtch2024]. People are not asking fewer questions. They are simply no longer asking each other. I have a personal stake in this story. My doctoral research at the University of Warwick examined helping as a form of participation in an open online community of independent game developers. I spent years watching how strangers helped strangers: why they did it, how they did it, and what it cost everyone involved [@hanif2016]. So allow me to say something that both sides of this debate tend to miss. The help was never free. One of the central findings of my thesis was that helping in online communities comes with conditions. Help was shaped by norms you had to learn, hierarchies you had to navigate, and reputations you had to earn. Ask a question the "wrong" way and you might be ignored, redirected to a search engine, or publicly scolded for not reading the documentation. Newcomers paid a social tax before they received an answer. The helpers were generous, yes, but their generosity operated within an economy of status, reciprocity, and belonging. Seen through that lens, the migration to AI makes perfect sense. The machine answers at two in the morning. It never sneers, never marks your question as duplicate, never makes you feel small for not knowing. It optimises your time and spares you the friction of human moods. For a working developer, a stressed student, or a teacher preparing tomorrow's lesson, this is not laziness. It is a rational trade. But every trade has a price, and we are only beginning to see this one. The price is that the conditions I documented were not only barriers. They were also the connective tissue of community. The social tax that newcomers paid was, in another light, an initiation. Answering questions was how strangers became members, how members built reputations, and how communities accumulated a shared, searchable body of knowledge that benefited millions who never posted at all. Every question answered privately by an AI is a question whose answer never enters the commons. The researchers behind the Stack Overflow study make this point starkly: when we substitute machines for public knowledge sharing, we shrink the very pool of human-generated knowledge that future systems, and future people, will learn from [@delriochanona2024]. There is a deeper cost still. Putnam's warning was never really about bowling; it was about what happens to a society when its habits of association wither. The Internet is now rehearsing its own version of that decline. Interaction is increasingly reduced to individuals conversing privately with an algorithm, then sharing the polished outputs with one another. We are becoming broadcasters of AI-generated artefacts rather than participants in each other's problems. Tellingly, Burtch and colleagues found that Reddit's developer communities, where social bonds are stronger and interaction is not purely transactional, showed no comparable decline. Social fabric, it turns out, is a buffer against the community-degrading effects of these tools [@burtch2024]. The communities that existed only to exchange information are dying first. The ones built on relationships are holding on. Putnam could have predicted exactly that. And there are early signs that the individual, not just the community, pays something too. Recent research involving 666 participants found that heavier reliance on AI tools was associated with weaker critical thinking, mediated by what psychologists call cognitive offloading, with younger users showing the strongest dependence [@gerlich2025]. When we stop wrestling with problems in front of other people, we may also be wrestling less, full stop. I will be honest about where I land. I do not believe this trend can be reversed, and I am suspicious of anyone who claims otherwise. The efficiency gains are real, the convenience is real, and no appeal to nostalgia will persuade a generation to queue for answers that a machine provides instantly. This is the price of advancement, and history suggests we always pay it. We did not un-invent the car to save the kampung's walking culture. But accepting a trade is different from ignoring its terms. Because this scenario is now arriving in our classrooms. Students who once asked their friends, their seniors, or their lecturers are asking a chatbot instead. The struggling student who would have been noticed in a study group now struggles alone, efficiently. The peer-to-peer helping culture that educators spent decades cultivating, the very behaviour I studied among strangers on the Internet, is being quietly outsourced. So the question I leave for my fellow educators is not whether AI belongs in education. That question has been answered by our students, whether we like the answer or not. The question is what we deliberately preserve. If helping one another is no longer necessary, it must become intentional: designed into our courses, our staffrooms, and our platforms, because it will no longer happen on its own. Putnam feared a nation bowling alone. Our students may soon be learning alone, in rooms full of people, each with a perfectly helpful machine. Are we ready for that? ## References - Burtch, G., Lee, D., & Chen, Z. (2024). The consequences of generative AI for online knowledge communities. Scientific Reports, 14(1). https://doi.org/10.1038/s41598-024-61221-0 - del Rio-Chanona, R. M., Laurentsyeva, N., & Wachs, J. (2024). Large language models reduce public knowledge sharing on online Q&A platforms. PNAS Nexus, 3(9). https://doi.org/10.1093/pnasnexus/pgae400 - Gerlich, M. (2025). AI Tools in Society: Impacts on Cognitive Offloading and the Future of Critical Thinking. Societies, 15(1), 6. https://doi.org/10.3390/soc15010006 - Hanif, H., & Hammond, M. (2016). Why and how do members provide help for others within online communities? Proceedings of the International Conference on Networked Learning, 10, 385–389. https://journals.aau.dk/index.php/nlc/article/download/8896/7152 - Putnam, R. D. (2000). Bowling alone: The collapse and revival of American community. Simon & Schuster. --- Canonical: https://hackedu.tech/articles/prompting-alone-what-we-lose-when-we-stop-asking-each-other-for-help Licence: CC-BY-4.0