Synthetic profiles, plausible faces

Content creators generated with artificial intelligence are entering a different phase: they no longer appear only as obvious technological curiosities or easily identifiable digital characters, but as profiles increasingly similar to any human influencer appearing on Instagram, TikTok, or X. A new report from *The Verge* describes how that shift is making it harder to distinguish whether behind an account is a real person, a creative agency, or an automated generation pipeline designed to capture attention and monetize it.

During the first wave of virtual influencers, the best-known examples looked almost like public demonstrations of the technology. Figures such as Lil Miquela, Imma, and Shudu Gram retained a deliberately artificial aesthetic and were presented fairly clearly as digital products. Now, according to *The Verge*, newer accounts such as Emily Pellegrini or Aitana Lopez occupy much more ambiguous territory: travel photos, restaurants, events, lifestyle poses, and posts that no longer look like futuristic exceptions, but like part of the normal stream of aspirational content dominating the platforms.

What is really changing

The novelty is not just that generated images have improved. What is changing is the scale and the context. More accessible tools, lower costs, and faster production workflows are allowing more people and companies to create synthetic identities with a consistent appearance, a steady posting rhythm, and enough realism to go unnoticed in a saturated feed. *The Verge* maintains that these accounts are already part of a broader mass of automated content: posts copied from chatbots, generated videos, images made to exploit trends, and profiles created to sell products, attract traffic, or exploit emotional and sexual niches.

That makes it harder to measure the phenomenon's actual size. Platforms do not publish clear figures on how many of these profiles are actively operating, how much engagement they generate, or what proportion of visible content comes from synthetic or semi-synthetic systems. Databases such as Virtual Humans can track some of the most prominent avatars, but beneath that visible layer there is already a larger ecosystem of small, imitative, and less detectable accounts.

The problem is no longer just spotting a fake photo

The discussion has also shifted to another level. Previously, the focus was on whether an image had enough flaws to give away that it was fake. Now the question is broader: what happens when entire accounts, complete with personality, aesthetics, and a posting routine, can be manufactured to look authentic? In that scenario, the fabrication stops being an isolated piece of content and becomes a system of continuous presence.

That deterioration in trust fits with broader warnings about the digital information environment. In January, *NBC News* reported that AI is intensifying a “collapse” of online trust, in a context where fake images, altered videos, and recycled material already mix with real events during moments of intense public attention. The problem is not only that convincing deepfakes exist, but that the experience of browsing social media is beginning to be marked by constant doubt about what is real, what has been edited, and what was fabricated from scratch.

Academic research suggests that this problem is not solved by labels alone. A study published in *Communications Psychology* found that even when people receive an explicit warning that a video is a deepfake, that content can continue to influence their moral judgments. The finding does not refer directly to virtual influencers, but it does reinforce an important conclusion for this story: once synthetic content achieves sufficient realism and circulation, transparency alone may not be enough to neutralize its effect.

Platforms, business, and moderation

For platforms, this opens up an uncomfortable challenge. The social media business model rewards frequency, visibility, and engagement, not deep authenticity. If a synthetic identity can produce attractive, consistent, and inexpensive content, the commercial temptation is obvious. At the same time, the cost to public trust also grows: fake profiles can sell products, manipulate audiences, simulate intimacy, or push misleading narratives without having to look like obvious robots.

The story, therefore, is not just about aesthetics or internet culture. It also touches on moderation, transparency, commercial responsibility, and information security. The underlying question is no longer whether AI-made creators will exist, but how they will be labeled, what limits platforms will impose, and how long it will take the average user to accept that a significant portion of their feed may not come from real people.

An increasingly ambiguous social media landscape

What *The Verge* documents is a quiet but structural shift: social networks no longer host only content created by humans with AI assistance, but also entire digital identities that can compete for attention on terms increasingly similar to those of any traditional creator. If that trend continues to grow, the central problem in the coming years will not be only the quality of fake images, but the erosion of the basic trust that sustained the social experience on the internet.