- Generative AI may have unintentionally pushed out some of the highest quality “expert” contributors on sites like Stack Overflow, as users instead increasingly adopted tools trained on their comments.
- The problem is that users feel like their expertise and efforts are not being rewarded, with AI often delivering the same solutions at a faster pace.
- This development is not limited to online coding communities, but threatens to expand to other areas such as classrooms, corporate workplaces, and scientific communities.
A University of Auckland study into the demise of Stack Overflow in recent years reveals an increasingly worrying trend in the software community: the best, or most knowledgeable, contributors are leaving in droves.
AI, which arguably bridges the gap between most entry- and mid-range coders and some of the best in the business, could actually accelerate the latter’s exit from online communities as they feel their efforts are no longer as valued as they once were.
Stack Overflow has seen a nearly 76% decline in monthly questions posted since the advent of ChatGPT in 2022, indicating that new and existing users are abandoning the site.
A much bigger problem than just stack overflow?
Stack Overflow’s problems and reasons for its decline were multiple; however, many users felt that the site and some of its most talented contributors displayed a degree of hubris.
This, coupled with authoritarian moderation that many describe as “self-righteous”, meant that users finding a viable option would inevitably leave the platform.
ChatGPT and its AI alternatives became considerably more flexible and, over time, became search engines for many coders with routine, repeatable queries, even as AI handled issues such as syntax issues increasingly better than before.
This, in turn, reduced the number of questions asked on the platform and, despite a ban on generative AI enacted shortly after ChatGPT went live, led to a loss of respondents that may prove impossible to replace in the long term.
The problem may no longer be limited to online coding communities; the researchers say this could spread to other domains such as classrooms, offices, and other research communities, where low-effort responses are more difficult to distinguish from those of subject matter experts thanks to retrained and constantly evolving AI models.
“If everyone can create a good quality response or result using AI, some people might think, ‘Why should I make an effort to share my expertise and participate?'” explained study editor Dr. Kenny Ching.
Ching called this “signal compression” because expert and non-expert solutions became harder to separate, even as it became less rewarding to be a subject matter expert on topics that AI could also easily intervene on.
The question that comes to mind here, however, is simpler: If AI was trained on user-contributed data and a smaller and smaller amount of it existed on platforms like Stack Overflow, where will the next knowledge reset take us in terms of AI’s capabilities?
While future AI models won’t become “dumber,” so to speak, they might turn to different training avenues, such as Slack chats, Discord servers, or even users currently asking them the same coding-related questions they once asked on Stack Overflow.
Whether this replaces experts who no longer wish to contribute or simply makes AI more error-prone over time, through the operation of its feedback loop, is an interesting question in a society that finds it increasingly difficult to distinguish between AI and human responses.
Follow TechRadar on Google News And add us as your favorite source to get our news, reviews and expert opinions in your feeds.
