Think back to the most recent movie you saw that you really liked. What did you like about it? What sort of impression did it leave?

Now, think back to one of your all time favorite movies — one you always like — a movie you’d watch again and again, anytime. What keeps you coming back? What makes this movie so special to you?

Now let’s consider some key differences between each of these movies.

Which one pops up in your mind’s eye when you’re just moving through your day? Which one resonates more deeply on an emotional level? Which one have you recommended to friends more?

The difference between the ways you responded to these questions surfaces some of the qualities that distinguish meaningful from compelling content; and this difference is becoming increasingly critical in what we now seem to be calling the ‘attention economy.’ Most importantly, this difference is a core part of determining how algorithms are shaped, which plays a critical role in establishing and reinforcing what constitutes ‘common’ knowledge (think content-scraping for training AI models), as well as which content becomes resilient (since the role of memory is an essential part of this difference — more on that later).

To get us started, I’ll explore how both meaningful and compelling content have played distinctly different roles as the web matured. Then, I’ll consider how we interact with each, the value ascribed to them, and how they’ve been (and could be) reconfigured as we head into a new era of interactions in the age of AI.

Breaking it Down

A close look at the ways the tech industry traditionally related to content helps set the stage. Although the history is diverse, the industry originally focused on the informational and the factual. Content was, simply, data. This made sense, given the core nature of digital technology itself, and the way programmers and engineers drove most of how tech products were developed. From strings of zeros and ones to coding logics, these systems were ultimately rooted in binary ‘call-and-response’ interaction models: inputs and outputs. Content functioned technically as a resource that could be referenced or processed. Eventually, the internet took this model to another level. It made information available globally through a robust and decentralized distribution system that had never existed before. Yet, the interaction models within it remained relatively constrained by utilitarian transactions.

Then came social media. While you could argue that it was originally conceived in primarily transactional terms — finding and ‘friending’ others — it quickly became evident that many users found these interactions to be valuable beyond the transactional. They valued the connection itself. As these platforms matured, there were clear indications that people wanted to include a much broader range of expressions, interactions, and co-modifications of content that included qualities like nostalgia, commemoration, fondness, flirting, and humor (to name a few).

It was perhaps no accident that the rapid growth of social media nested comfortably within a parallel rise of reality entertainment (and the decline of traditional media institutions). There’s much more to consider here, including the long (and deep) tradition of seeing ourselves as sources of entertainment (parlor tricks, seances, folk dance, ice bucket challenges, etc.) — more on that in another post. The point I want to make here is that transactional relationships with digital content only went so far. People were seeking deeper engagement, with content and with others. Their impulses were driving them to look for much more than just getting an answer. They were seeking exchanges with other humans, channeled through content that was emotional and authentic. We’re still seeing this play out in our digital lives everyday, of course.

Whether you perceive these emotions and connections as positive influences on our collective social fabric or not, it’s clear that the range of our interactions has expanded, and our shift from transactional to social revolutionized the way we think about digital content. A techno-centric view might posit that technologies like social media ushered in a new wave of behaviors. An anthropological one might argue that those behaviors always existed, and the technology caught up to (and exploited?) them.

“Huh” or “Hell, Yes!”

Let’s return to the differences between meaningful and compelling. While meaningful experiences are typically tangible, useful, and timely, they don’t necessarily fill us with the urge to reach out and share with others. Their value comes in the form of knowledge and understanding. In contrast, compelling interactions are magnetic, personal, and tap shared cultural narratives. Even more importantly, compelling interactions are catalyzing. We’re drawn to them again and again because they resonate so deeply for us. This depth propels ownership, action, and agency.

The table below summarizes some of the core differences I’d like to highlight.

Meaningful	Compelling
Transactional / Informational Tangible Useful Timely	Engaging / Magnetic Deeply Personal Culturally Relevant Catalyzing
Outcome: Knowledge & Understanding	Outcome: Ownership, Action & Agency

In other words, meaningful content makes you go “huh,” or maybe “I see.” Whereas compelling content makes you go “OMG, did you see THIS?!” Meaningful content helps you feel informed, appreciative, or even satiated, but its function is ultimately utilitarian — then you move on. With compelling content, you feel driven to share, to build on the momentum of your enthusiasm by adding your own energy to it (sometimes repeatedly).

Value and Risk

In terms of our daily lives and interactions with content, we can think of meaningful content as the something that’s often used to resolve small disagreements, to navigate, to make decisions, or to learn something new. Overall, it’s a tool, not a propellant. Compelling content, on the other hand, has a distinctly emotional component. Interacting with it is embedded with qualities we don’t associate with the transactional nature of meaningful content.

Psychiatrist Dr. Goulston argues that when people are engaged in compelling interactions, they feel respected, engaged, and invited. They feel talked with (instead of talked at, or over), and then behave as if they’re choosing to do something. Perhaps most importantly, compelling interactions catalyze the initiation of new interactions from people exposed to content they find deeply engaging and highly resonant. They’re driven to share their experiences with others (remember your all time favorite movie?).

In this light, it’s easy to see how compelling content can set the stage for increased customer loyalty and brand resonance in marketplace settings. Examples include engagements with content that grow organically within a culture or community of customers. Emotional connection, authenticity, and active participation are some of its core characteristics. Here we’ll find superfans (not paid ‘brand ambassadors’) who enthusiastically share content that’s compelling to them. This can be exceptionally powerful and empowering when organizations are savvy enough to include them. They co-create, co-opt, and co-evolve brand identities (think IKEA hacks, cosplay, or fan fiction).

It’s also easy how to see how compelling content can be misused. In Jonathan Haight’s now classic analysis of moral judgments, he uses the metaphor of a rider sitting atop an elephant to illustrate how we reach moral decisions. His argument is that intuition (the elephant moving along) tends to come first, dominating our moral compass. Its nature is to go where it likes (what compels it). Rational thinking, represented by the rider, has some control, but is secondary to the often overpowering drive of the elephant.

Hacking this dynamic by ‘luring’ the elephant with compelling content is the risk I’m referring to here. There’s now ample evidence that conflict draws clicks, as does other content that one could argue has questionable social value. But the intent isn’t always nefarious.

Motivational science also plays on this dynamic to optimize for positive behavior outcomes. For example, in the early days of recycling in the US, it was initially difficult to persuade people to change well-engrained behaviors around waste disposal. Gradually, however, municipalities began to realize that if they tethered their recycling goals to something people cared deeply about — their children — they could influence household habits. Many launched recycling education campaigns in schools, which included both take-home materials and a new environmental consciousness students brought home to their parents, who were then ‘compelled’ to recycle at least in part out of their deeply rooted love and care for their children. Ownership, action, agency.

Beyond Binary: AI

What does this difference between meaningful and compelling mean for the digital world, and especially the web, moving forward? One view is that meaningful content may increasingly be subject to automated and endless reinterpretations via AI’s ability to scrape, summarize, re-word, and translate content. This could lead to commonplace repackaging of meaningful content that personalizes it in ways that give it ‘voice’ (both literally and figuratively). Could this transform meaningful content into compelling content? Will AI essentially merge the two in the form of ‘companions’ (or agents) that time, contextualize, and find personal points of relation and activation of meaningful content for us?

This potential merger (dissolution’?) is something our collective imaginations have only recently entertained. Even in the fairly recent past, our interpretation of how we might interact with AI was constrained well within the bounds of traditional meaningful-content conceptualizations. Take this clip from the film AI: Artificial Intelligence, where AI is portrayed as a font of knowledge largely conceived of as a giant vending machine for human queries. Here we see a future in which meaningful content is simply scaled up, but not reinvented as compelling content. In fact, while Dr. Know is all-knowing, his services are provided purely on his terms — and he’s quite literal!

We’ve come a long way in thinking about how we might interact with computers, in a short amount of time.

If we take the potential dissolution of distinctions between meaningful and compelling interactions and extrapolate it, it’s not hard to conceive of a post-meaning AI-dominant online world, where compelling content and interactions are ALL that matter — and meaning (or accuracy) is served up as a secondary priority embedded within compelling forms, tailored to our interests and preferences. Some would say this is already happening. They point toward common hallucinations and inaccuracies that ‘sound right,’ and are positioned by AI as correct, but are completely fabricated based on familiar patterns the tools have identified. In short, AI models may inherently index toward compelling at the expense of meaning (or accuracy). Another risk is that institutions or organizations generating AI content may use this to leverage our ‘elephants’ in ways that obfuscate the reasoning of our ‘riders.’

In a recent piece that explores the impact of AI hallucinations on our interactions with content, Matteo Wong takes this a step further: “AI products could settle into a liminal zone. They may not be wrong frequently enough to be jettisoned, but they also may not be wrong rarely enough to ever be fully trusted. For now, the technology’s flaws are readily detected and corrected. But as people become more and more accustomed to AI in their life—at school, at work, at home—they may cease to notice. Already, a growing body of research correlates persistent use of AI with a drop in critical thinking; humans become reliant on AI and unwilling, perhaps unable, to verify its work.”

Maybe. Or we could see an erosion of trust in online content (including suspicion that its purveyors are toying with our ‘elephants’) that leads to a rise in reliance on verifiable real life experiences. Events, gatherings, and any other interactions not mediated digitally may grow in importance, as a sort of radical empiricism rises. These gatherings (be they salons, conferences, clubs, debates, or other) may become the only way we feel we can find what we consider sincere, faithful, meaningful interactions, and authentically-expressed enthusiasm for compelling content. This is not to say that AI couldn’t play a useful role in organizing such events, or identifying alignments between them and potential attendees, but that ultimately the value of real life gatherings may lie in our collective appreciation for face-to-face experiences in which meaning is generated through conversations, and compelling interactions are ascertained by looking directly into some one’s eyes.

There’s an important take-away here for Silicon Valley: it’s probably not a good idea to throw AI at every possible interaction. Product-market fit isn’t rooted in what’s technically feasible, but in the value and trust a product builds with its users. At a minimum, considerations for using AI in any product should include the degree to which users are seeking meaningful versus compelling content and interactions, as well as the potential risk for inaccuracies to erode trust. This may not matter much in creative tasks, or conversational interactions; but when we need to rely on content and value its meaningfulness for productivity, for example, the stakes are much higher.

In Closing: A Break from it All

I’ll leave you with the clip below from a recent stay in New York — a glimpse of people engaging in playful acts of joy, scooting around in bumper cars in a rink of giant pink inflated balls. No phones. Probably no meaning. Just deeply compelling fun. And, perhaps most notably, no AI.