The Generative AI Impact on Online Communities: Are Chatbots Closing the Door on Beginners?
Source PublicationScientific Publication
Primary AuthorsDu, Li, Yu et al.
"Imagine a bustling local pub where newcomers used to ask regulars for directions. If someone installs a digital map by the door, beginners get their answers and leave without ever chatting with the locals. Eventually, the pub just becomes a closed circle of old friends."
Imagine a bustling, healthy coral reef. It thrives on tiny, everyday interactions. Small fish nibble on algae to clean the coral. The clean coral then provides a safe harbour for larger fish. Everything is connected. But what happens if you drop a massive, artificial food dispenser into the water? The small fish get their meals easily. They stop visiting the coral. The big fish might stay in their usual spots, but the reef stops growing. It becomes a closed-off fortress, slowly losing its vibrant flow of new life.
This biological safehouse is a perfect way to understand how we share knowledge on the internet.
The Generative AI Impact on Online Communities
For years, digital forums like Stack Exchange have functioned just like a reef. Beginners arrive with basic questions. Experts answer them. Through these simple interactions, newcomers learn the rules, make friends, and eventually grow into experts themselves. Scientists call this process legitimate peripheral participation. It is how groups renew themselves.
However, recent research suggests this natural cycle is breaking down. A new study measured the generative AI impact on online communities by tracking seven Stack Exchange forums between 2016 and 2025. The researchers treated the public release of ChatGPT as a sudden shock to the system. They wanted to see what happens when an algorithm can answer basic questions instantly.
Measuring the Missing Beginners
The findings paint a stark picture. In forums focused on computer programming and algorithms, the number of newcomers who stayed for at least a year fell by nearly half.
Let us break down exactly how this works step-by-step. First, a student encounters a problem with their computer programme. In the past, they would create an account on a forum to ask for help. Second, a veteran programmer would reply, perhaps offering a tip on formatting alongside the answer. Third, the student would thank them, feeling a small sense of belonging. Over time, that student would start answering questions for others.
Today, that first step is intercepted. If a chatbot can write your code, then you do not need to ask a human for a favour. You get your answer in seconds and close the tab. Because of this behaviour, the study measured a sharp decline in low-activity members joining and staying in the network. Meanwhile, the retention of highly active, veteran members remained completely stable. The AI acts as a substitute for information, but it cannot replace physical experience, which is why practice-based forums saw less disruption.
A Fractured Social Structure
What does this mean for the future of our digital spaces? The data suggests that communities are becoming more closed off. Without beginners asking questions, the bridges between different groups of experts begin to weaken. The network becomes fragmented into isolated bubbles.
If we rely entirely on chatbots for our routine questions, then we lose the entry-level interactions that build human connections. While the study could not directly observe individual decisions to use an AI, the structural changes are clear. The old guard remains bonded, but the door for new members is slowly closing. The reef is still alive, but without the small fish, its future looks very different.