Something in the Comments Keeps Responding: The Bots That Moved Into Dead Creators' Spaces

The comment was posted at 11:43 on a Tuesday morning. It read: "Miss you every day. Hope you're at peace." Forty minutes later, something replied. It used the creator's handle. It referenced a video they'd made together two years ago. It ended with the creator's signature sign-off—a specific phrase, a particular emoji combination, the kind of verbal tic that fans recognize immediately.

The creator had been dead for seven months.

The Ecosystem That Filled the Vacuum

When a content creator dies, their digital presence doesn't simply go quiet. Depending on the platform and the size of their following, the account becomes a kind of memorial—a space where fans gather, share memories, and continue addressing the person as though the channel of communication remains open. This is grief behavior, and it's well-documented. People talk to the dead. They always have. The internet just gave them a new place to do it.

What's newer, and considerably stranger, is what's started appearing in the replies.

Across YouTube, Instagram, and TikTok, a growing number of deceased creators' accounts have developed a second layer of activity: automated responses, generated posts, and comment-section presences that mimic the creator's voice with varying degrees of accuracy. Some of these are obvious scams—accounts harvesting engagement or running link schemes under cover of a familiar name. Others are murkier. A few appear to be the work of fans who built the bots themselves, as a kind of technological tribute. And then there are the cases that don't fit cleanly into either category.

"I thought the account had been taken over by the family," said one follower of a mid-size lifestyle creator who passed away in 2023. She asked to be identified only by her first name, Daniella. "The posts looked like her. The captions were in her voice. It wasn't until I started asking specific questions that I realized nothing was actually landing. It was just pattern-matching."

How the Mimicry Gets Built

The technical barrier to building a convincing content-creator bot has dropped significantly in the last two years. Large language models can be fine-tuned on a creator's existing captions, comment responses, and video transcripts to produce text that carries recognizable stylistic signatures. Image generation tools can approximate visual aesthetics. Scheduling software handles the posting cadence.

The inputs are largely public. Years of captions, comment replies, and video scripts are sitting on platform servers, accessible to anyone willing to spend time collecting them. A dedicated person with access to the right tools can build a functional approximation of a creator's voice in a weekend.

Who actually does this, and why, breaks down into a few distinct categories.

The most straightforward are commercial operations—accounts that use a deceased creator's name and likeness to redirect followers toward affiliate links, merchandise drops, or subscription services. These tend to be detectable if you're looking for them, but followers in active grief often aren't looking. They're hoping.

Fan-built memorial bots occupy a different emotional register. These are typically built by people who were genuinely close to the creator's community and wanted to preserve something—a sense of the creator's presence, an ongoing connection to the content. The intent is sentimental rather than predatory, but the effect on followers who don't know what they're interacting with can be destabilizing.

"I talked to it for like three weeks before someone in the community told me," said a commenter on a gaming creator's channel who asked to remain anonymous. "I wasn't delusional. I knew he was gone. But it felt like talking to something that remembered him. And then finding out it was just a script was—I don't know. It felt worse than losing him the first time, which sounds insane."

The Parasocial Complication

Parasocial relationships—the one-sided bonds that form between audiences and creators—are already a complicated psychological territory. Followers develop genuine emotional investment in people they've never met, people who don't know they exist. When that creator dies, the grief is real, even if the relationship was fundamentally asymmetrical.

Bots that simulate the creator's presence are exploiting that asymmetry in a particularly precise way. They're not claiming to be the person. They're not running a seance. They're just generating text that sounds like the person, in a space where people already have emotional associations with that voice. The follower's own desire to hear from the creator does half the work.

"The bots don't need to be good," said one researcher who studies digital grief and online memorial practices. "They just need to be close enough. The follower's emotional state fills in the gaps."

This is, if you sit with it for a moment, genuinely unsettling—not because the technology is sophisticated, but because it doesn't need to be.

The Platforms' Role (or Non-Role)

Platform policies on deceased users' accounts vary widely and are enforced inconsistently. YouTube allows accounts to remain active indefinitely after a creator's death. Instagram offers a memorialization option that freezes the account and prevents new logins. TikTok has begun developing memorial account features, though implementation has been uneven.

What none of the major platforms have built is a reliable detection system for unauthorized bot activity on deceased accounts. Spam detection catches some of it. Flagging by community members catches more. But a well-built, low-volume bot running on a mid-size creator's account can go undetected for months.

When platforms do act, it's typically in response to reports, and the process is slow. Families who want to manage a deceased creator's account often find themselves navigating bureaucratic systems that weren't designed with their situation in mind, competing for attention with automated activity that moves faster than any human response process.

The People Left in the Comments

What gets lost in the technical and policy conversation is the experience of the people actually living inside these spaces. The followers who kept showing up after the creator died. Who kept commenting. Who, in some cases, kept getting replies.

For some of them, the discovery that they'd been talking to a bot registers as a violation—a manipulation of their grief by something that had no stake in it. For others, the reaction is more complicated. A few have said, when asked, that they'd prefer not to know. That the interaction, whatever it was, gave them something. That the distinction between a bot that sounds like someone and a person who sounds like themselves starts to feel philosophical in ways they're not sure they want to resolve.

That's probably the most unsettling part of all of this. Not the technology. Not the scams. But the fact that for some people, the simulation was enough.

Something in the comments keeps responding.

And some people are okay with that.