Social Media, AI, and the New Reality of Influence

Originally published in IASIU's SIU Today Summer 2026 Newsletter

A PREDICTION THAT QUIETLY BECAME REALITY  

Back in 2023, I wrote about something that, at the time, felt like a trend that was still forming: the rise of the influencer economy. It was framed as a shift in marketing, engagement, and online behavior, but the underlying message was simpler: social media was no longer just a place where people shared their lives. It was becoming a place where people performed them. In that earlier piece, I noted that the influencer economy would continue to grow into a dominant force, shaping how consumers engage with brands, products, and even personal identity. At the time, it was easy to read that as commentary on advertising or digital culture. Looking back now in 2026, it reads more like a transition point, not a prediction of growth, but a description of structural change. Because what we are seeing today is no longer just the rise of influencers. It is the replacement of the original social media experience entirely.

THE QUIET SHIFT NO ONE REALLY CHOSE

For years, social media platforms were built on a simple promise: you would connect with people you know. Friends, family, coworkers, your digital life was meant to mirror your real one. That is no longer the dominant experience. Today, most users open their feeds and are met not with updates from their personal networks, but with creators, individuals they do not know, producing highly optimized content designed to hold attention for as long as possible. That shift was captured recently by Mark Zuckerberg, who described it plainly: “Social media started out as people primarily interacting with their friends. And now… at least half of the content is basically people interacting with creators.” It is a simple observation, but one that reflects a profound structural change in how information flows through modern platforms.  

“The value of the data being created often far exceeds the user’s understanding of how it may be used.”  

We did not collectively decide to replace our social circles with content creators. There was no single moment of transition. Instead, the algorithm quietly tested, optimized, and adjusted, and over time, it learned something very important: Strangers outperform friends. Not emotionally. Not meaningfully. But in engagement, in watch time, clicks, retention, and interaction. And in an attention-driven ecosystem, that is the only metric that matters.

FROM INFLUENCERS TO SYNTHETIC INFLUENCE

If the last few years were defined by the rise of the influencer economy, the current phase is defined by something more complex: the merging of social media and artificial intelligence. We are now entering an environment where content is no longer just created, it is generated, optimized, and iterated at scale. The feed is no longer simply curated by humans or even traditional algorithms. It is increasingly shaped by systems capable of producing infinite variations of highly engaging content tailored to individual users. The result is a shift that is subtle but significant. Influence is no longer just about reach or personality. It is about optimization. Content that performs well is rewarded. Content that does not is buried. And increasingly, the distinction between human generated and AI-generated material is becoming irrelevant to the platform itself, so long as the engagement metrics remain strong. This raises an uncomfortable question for those of us working in investigations, insurance, and fraud analysis: If content can be generated to perfectly mirror human behavior, what exactly are we validating when we say something “looks real”?

THE INVESTIGATIVE IMPLICATIONS

For SIU professionals, this evolution is not theoretical. It is already impacting the way digital evidence is created, shared, and interpreted. We are entering a space where:

• A single individual can maintain multiple digital identities across platforms with minimal friction
• Images, videos, and even “live” content can be synthetically generated or enhanced
• Behavioral patterns can be manufactured to support a narrative over time
• Traditional assumptions about authenticity no longer hold  

The challenge is no longer simply identifying what someone posted. It is determining whether the person behind the post, or even the post itself, is part of a constructed system of influence.  

“What feels like entertainment on the surface may actually be infrastructure underneath.”  

This is where the work becomes more complex, and arguably more important. Because as digital environments become more synthetic, the burden on investigators to validate truth increases significantly.

A WARNING THAT AGED QUICKLY

In 2023, I wrote that artificial intelligence would become increasingly embedded in everyday life and in the claims environment. At the time, the statement was intended as a cautionary note, a reminder that AI was not a passing trend, but a structural shift that would need to be understood, not ignored. That statement has proven accurate far faster than most anticipated. AI is no longer emerging. It is operational. It is embedded in content creation, recommendation systems, fraud generation, identity manipulation, and even in how people construct their digital presence. What was once a tool is now an environment. And within that environment, the line between authentic and artificial is becoming increasingly difficult to draw using traditional methods alone.

THE HIDDEN VALUE OF USER-GENERATED DATA

A recent viral discussion surrounding Pokémon Go highlights a broader truth about the modern digital ecosystem. While players believed they were simply interacting with a game, they were also contributing to a massive, real-world visual dataset through location-based gameplay and AR scanning features.  

Companies like Niantic have openly acknowledged building largescale spatial maps using user-generated imagery, not as a hidden operation, but as a core part of their augmented reality platform.  

The takeaway is not that users were misled, but that the value of the data being created often far exceeds the user’s immediate understanding of how it may be used.  

For investigators, this reinforces a critical point: digital activity is rarely one-dimensional.  

What appears to be entertainment on the surface may also serve as infrastructure, data that can be leveraged for mapping, modeling, or artificial intelligence development.

RESPONSIBILITY IN AN AI-DRIVEN ECOSYSTEM

This leads to a broader issue that extends beyond investigations or insurance: responsibility. There are two layers to this. The first is personal. Every interaction online, every view, like, share, and comment, contributes to training the systems that now determine what we see. In many ways, users are not just consuming content; they are shaping the systems that decide what content exists.  

“Players weren’t just catching Pokémon, they were mapping the real world, one scan at a time.”

The second is professional. For those in SIU, claims, and fraud prevention, the responsibility is to adapt investigative frameworks to an environment where authenticity can no longer be assumed by default. That means:  

• Moving from passive observation to active verification
• Treating digital content as potentially synthetic until validated
• Strengthening documentation, chain of custody, and metadata preservation Understanding that the absence of evidence is no longer meaningful in isolation  

The investigative standard has not changed, but the environment in which it applied has changed.

CLOSING PERSPECTIVE

In 2023, the conversation was about influencers. In 2026, the conversation is about something more fundamental: the nature of influence itself. We are no longer simply observing social media. We are operating inside a system where attention is engineered, identity is fluid, and content is increasingly indistinguishable from simulation.

“No fleet of cameras could replicate what millions of people casually create every day.”

The shift did not happen all at once. It happened gradually, almost invisibly, one optimized feed at a time. And perhaps that is the most important point of all. The most significant changes are rarely the ones that announce themselves. They are the ones that simply become normal. Because the next evolution isn’t coming. It’s already here.

‍