AI Infiltrates Online Communities: Users Feel Threatened, Platforms Grapple for Balance

13

In the thriving, bustling hub that is our digital universe, every comment posted on Reddit, every code-related query solved on Stack Overflow, every touch-up contributed to a Wikipedia entry, even the intimate pictures shared on your public Facebook or Instagram feeds – each is quietly informing, teaching, and moulding artificial intelligence’s next generation. While it may seem innocent, it’s a cause of concern for many.

This apprehension stems partly from seeing the very same online forums they’ve invested their time and intellect into, becoming inundated with commentary conjured not by fellow humans, but artificial intelligence. It’s an eerily accurate mimicry, a digital parrot trained on reproducing human thought, with some users scrambling to erase past contributions or scrambling them into senseless chatter – futile attempts to resist the inevitable AI takeover.

Follow us on Google News! ✔️


Even as governments, like Brazil’s privacy regulator, have intervened, their efforts seem to be droplets in this daunting AI ocean. Longtime Reddit user turned volunteer moderator, Sarah Gilbert, who also indulges her curiosity in online communities and their dynamics at Cornell University, stated candidly, “A significant slice of the population feels helpless.” The only recourse seems to be a digital detox, depriving themselves of the value they derive, and can offer online.

Platforms are responding, albeit with mixed success. Stack Overflow, a go-to site for computer programming tips, initially rebuffed ChatGPT, the AI-writing bot infamous for its errors, only to later embrace an alliance with AI developers and penalize rebellious users attempting to delete their past inputs. It’s an apt illustration of the tightrope that social media platforms now walk, juggling user skepticism while trying to acclimatize to the technological flux instigated by generative AI.

Andy Rotering, a software developer from Bloomington, Minnesota, who has religiously used Stack Overflow for fifteen years, voiced concerns that the company could unintentionally harm its most valuable asset – its dedicated, wellspring community that willingly donates time to assist other programmers.
Encouraging contributors to continue their commentary should be a priority, he maintains.

Prashanth Chandrasekar, CEO of Stack Overflow, elucidated the company’s challenge in meeting the escalating demand for instantaneous chatbot-generated programming aid while preserving the essence of community interaction where individuals seek recognition for their contributions. Fast forward five years and he predicts an internet overrun with mechanical content, a dearth of spaces hosting original, authentic human thought. He intends for Stack Overflow to buck this trend.

Comparing Stack Overflow’s struggle to case studies discussed at Harvard Business School, Chandrasekar reveals their battle for survival in the wake of disruptive technological change.

Historically, Stack Overflow served as a programmer’s bible, home to coding solutions that would be Googled, found, and implemented. High credibility scores of contributors— a potential CV enhancement— determined these answers. Today, a question posted to an AI chatbot trained on Stack Overflow’s exhaustive databases yields an immediate response.

When ChatGPT made its debut in late 2022, it threatened to eclipse Stack Overflow altogether. Chandrasekar’s response was to quickly mobilize a 40-person team, leading to the launch of Stack Overflow’s own specialized AI chatbot, Overflow AI. Partnerships with Google and OpenAI provided the latter with access to the platform’s Q&A archives for refining their AI models. Maria Roche, an assistant professor at Harvard Business School, opines that this strategic shift makes sense, but wonders if it wasn’t implemented a little too late.

When Stack Overflow announced its OpenAI partnership, some users attempted to erase their historical comments. The company’s response was swift, suspending these users as violation of terms stating that contributions to Stack Overflow were “perpetually and irrevocably licensed.”

With Tuesday’s prohibition of Meta Platforms from training AI models on Brazilians’ social media posts, Brazil’s data protection authority underscored the ongoing online tussle. Meta called this a backward step, insisting its practices upheld Brazilian laws, and criticized the lack of transparency in other organizations undertaking similar AI training.

Reddit’s strategy diverges, partnering with AI developers and setting guidelines for content access, rejecting wholesale hoarding by commercial entities oblivious to user rights and privacy. The partnerships were lucrative, taking Reddit public on the New York Stock Exchange and driving the company’s valuation to almost $9 billion.

While Reddit hasn’t disciplined protesting users– a difficult task given voluntary moderators’ significant influence– Sarah Gilbert voices concerns about the surge of machine-created posts flooding Reddit. Believing users visit for human interaction, not bot conversations, she worries that an inflow of bot responses could deter users.

In a poignant closing note, Gilbert reflects on the irony of AI chatbots learning from and threatening to displace the very human commenters they were trained on, a looming, existential risk to the online community.