Source: Gao, Zhang & Lysyakov (2025) — “Does Social Bot Help Socialize? Evidence from a Microblogging Platform,” Information Systems Research
There is something slightly funny about platforms using bots to make humans more social.
It feels like hosting a dinner party and hiring a robot to laugh at everyone’s jokes so the guests feel encouraged.
This paper studies exactly that kind of situation. A microblogging platform launched a platform-owned social bot called CommentRobot. The bot generated comments under posts, and the researchers studied whether that public bot interaction increased user engagement.
The short answer is yes.
Posts that received bot-generated comments saw more engagement. An online experiment confirmed the effect and showed that both bot identity and comment quality mattered. Good bot comments were attractive, relevant, and included social cues. In other words, people responded not just because something commented, but because the comment felt socially useful.
That is the encouraging part.
The awkward part is what happened next.
Receiving bot comments encouraged users to create more bot-related posts, but it did not necessarily increase overall posting activity. So the bot helped socialize at the post level, but it did not magically revive the whole platform.
This is where the finding becomes very practical.
AI can stimulate engagement, but engagement is not always the same as community. A bot comment can make a post feel less lonely. It can create a nudge, a cue, a little spark. But if users begin posting for the bot instead of for other people, the platform may get activity without the deeper social value it wanted.
That feels like a very 2026 problem: the metric improves, but the meaning is questionable.
My takeaway: social bots can help, but deployment strategy matters. The goal should not be “make the numbers move.” It should be to create interactions humans still find worth joining. Otherwise, everyone is just performing for the machine, which is a bit too on the nose.