Why Curated Community-Generated Content is the Best Growth Channel in the Age of AI

AI spam has taken over our online world. We are currently in what seems to be a massive spam-driven AI era. AI-generated newsletters overflow our inboxes, AI-beautified media features prominently on our social media feeds, and AI-generated articles populate our search engine results. It may be distracting, sometimes even entertaining, but ultimately not helpful, and thus spam. While it is a nightmare for people receiving them, it’s also likely to become a hell for marketers and creators who create and distribute these pieces of communication.

This is simply because, although this kind of AI-generated content might temporarily result in increased click rates, such content sets people’s defenses on a high-alert, low-trust automatic response system. Consequently, it becomes extremely difficult to have a real conversation and erodes trust in a brand. The response to this, ironically, is a pivot toward the vox populi and community-generated content. How did we get there? To understand that, it’s important to understand the environment of low trust we’re operating in.

This is something the large curators of our Internet experiences (Google, Meta, LinkedIn) realize. This has prompted them to shift their algorithms to distribute our attention rewards differently. Google’s HCU or LinkedIn/Meta’s focus on suggested content, for example, are steps in that direction. They may be criticized for being highly imperfect steps, but they are still decisive pivots. The question for most businesses, then, is this: what to prioritize once the dust settles?

ALSO READ: How to Survive Google’s Helpful Content Update (and Grow)

Building High-Trust Systems

Before the AI buzz took over, most venture capitalists, investors, and founders were bullish on the idea of community-generated content. They wanted to build online communities where people trusted each other (even if in a limited context), which engendered a sense of belonging. These are spaces where a company can build a relationship with its customers and audiences. They can leverage community-generated content to understand their needs, behaviors, and desires. Not only did communities help with driving current business growth, but they additionally acted as a moat (a defensible advantage) to help with future growth as well. 

However, the AI craze disrupted many businesses. However, businesses that had an actively engaged community had a longer time to comprehend and respond to it. It may appear counterintuitive, but community-led platforms are thriving in the age of AI. Why? Aren’t we all just supposed to find our answers on a chat screen? What makes community-generated content so different or special?

Some of the biggest winners of this disruption have been Reddit, Quora, and even LinkedIn. Here are some numbers for context: Reddit has seen its organic reach grow by 4x between July 2023 to February 2024, while Quora scaled 3.5x during the same duration. Similarly, LinkedIn has been generating  >30 million SEO visits per month purely on the back of collaborative articles.

ALSO READ: What is Synthetic Media? Understanding AI’s Transformative Role

Answer vs. Answers: Why Community-Generated Content Works

If you were to ask complicated scientific/linear thought questions to an AI chatbot, it would surely save you much time and effort. Though AI falls short in the case of experience-centric questions—what is it like to watch Aurora, or how do you go through with a pregnancy when you have a demanding job? It can’t really share stories or anecdotes about what it really feels like. Humane experiences vary very much, which almost never really has an exact answer. And here is where community-generated content wins the day. 

Communities are built around those intangibles—anecdotes and stories. From the limited legal precedents, we know of, thankfully, LLMs (Large Language Models) can’t really just copy-paste such stories within their chat responses (at least not without attribution).

It seems like testimonials, anecdotes, and experiences are off-limits for current AI LLMs. Though not for you and me. In which case, isn’t having a community a wonderful asset to build, curate, and distribute truly helpful content? The best part is that community members are mostly quite happy to contribute back to help others.

ALSO READ: How Does LLM Optimization Influence AI Search Results?

Enter Community-Generated Content

Well, if you have used LinkedIn anytime recently, you might have come across people answering AI-prompted questions. Most people don’t really know the questions are AI-generated, but because they have a chance at winning a “top-voice” badge, they are happy to play along. Similarly, people contribute to Reddit and Quora (no known AI questions) for karma points and just to help others out.

On top of this, all three have a layer of bot-prevention activity built in. It’s not 100% reliable, but people upvoting particular responses or staking their past karma points/professional credibility is a decently good proxy for trust.

And yes, it aligns very well with Google’s emphasis around the E-E-A-T (Experience, Expertise, Authoritativeness and Trustworthiness) framework, which makes community-generated content from experts really valuable. Additionally, as humans, we have always loved learning what other people (more specifically, people like us) are thinking about a particular thing. No wonder websites like TripAdvisor or Glassdoor attract millions of people every month.

Curation is Critical for User-Generated Content

As discussed earlier, despite all the technological advances, a good way to figure out whether it will rain soon or not is to look outside the window. Likewise, context and curation are important. Thus, while community-generated content is important, it is only helpful when all the pieces fall together and connect well.

Here is an example: One of the largest climate-tech communities, Work On Climate, generates hundreds of messages and conversations each day around different climate-action topics. Most conversations are valuable in limited contexts, but not all of them would perform well on social media or search engines. As a consequence, making all of them openly sharable isn’t very helpful.

As an active member of the community, however, one could see what topics and themes are repeated most often. Questions like what was it like to do a Climatebase fellowship? Or are Terra courses worth the return on investment? Questions that keep repeating every time a new member joins, questions whose answers evolve over time, but elements of which also stay constant.

Collecting Meaningful Responses From Community-Generated Content

Now, let’s say we have identified the underlying question theme: which climate course would be best for my context? That’s not a question whose comprehensive answer can be given by just one member/administrator. Surely, as the context varies, the answer varies. LLMs aren’t really well-suited for contextual questions. At the same time, it can’t be expected of community members to keep answering the same/similar questions with high enthusiasm.

Thus, one way to go about answering such questions is to build curated guides that answer a question with a mix of general observations and accompanying quotes from community members. The guide can be divided into sections and subsections that help build contextual knowledge on different themes.

Consequently, not only does one create really helpful content that’s loved by the community, but there are also many other benefits—social sharing, SEO, community engagement, and attracting new members and customers. Though the biggest benefit is that you build more trust. In a low-trust market, after all, winning trust pays compounding dividends. Additionally, this content style can’t be replicated by LLMs without getting lost in a sea of legal issues.

Precautions and Limitations of User-Generated Content

This is a relatively new format. And while LinkedIn has exploited it much using AI-based prompts, there’s still a lot of innovation and creativity that’s about to happen in this space.

Still, here are a few cases where this structure will not work to create desired results:

1. Democratization isn’t a replacement for expertise. Therefore, this content style is more suited for topics that require low/medium expertise.

2. Questions where the answer isn’t subjective.

3. When questionnaires lack depth. Thus, framing the right questions is important as the quality of answers depends heavily on the quality of the questions.

4. Doesn’t fit well with short deadlines; people take their own sweet time to respond, so one has to be patient.

Above all else, one has to be cognizant of respecting people’s privacy and asking them for their permission to share their opinions.

Why Community-Generated Content Delivers

Despite the precautions, this content format is great due to a great alignment between the interests of different organizations. People are happy to read because they get authentic, helpful first-hand information. Community members are happy to contribute because they get to showcase their expertise in a low-friction format (much less effort than writing a new blog). Google rewards such content because people spend time reading it, and with a few SEO tactics (H1 tags, keywords, schema, backlinks), it works better. Community managers/admins are happy because the community genuinely finds it helpful. And in general people are happy to give shout-outs/backlinks through their networks when they find something truly valuable.

The LLMs might not be very happy about it, but then, this isn’t ice cream; it is a growth strategy for the digital marketing industry that is coping with disruption. 

NOTE: The views expressed in this article are those of the author and not of Emeritus.