Meta had been developing personality-driven chatbots for its apps since its earlier planned “personas” rollout, framed as a way to increase engagement. This report extends that approach from conversational tools to accounts that can publish and circulate material alongside users.
The idea quickly met a practical constraint: Meta later removed viral AI profiles from an early experiment after backlash, highlighting the gap between automating activity and making synthetic participation acceptable to users.
First-order effects
Meta’s product and policy teams would need to define how AI-generated characters are represented, governed and allowed to distribute content on its platforms.
Users and creators would encounter a potentially larger pool of platform-generated posts and interactions, rather than AI being limited to on-demand chat experiences.
Second-order effects
The backlash to Meta’s early AI-profile experiment makes disclosure, labeling and enforcement central product requirements; otherwise synthetic accounts can undermine confidence in ordinary account activity.
A larger volume of generated posts would intensify competition for attention in feeds, raising the importance of Meta’s ranking controls and of creators’ ability to distinguish their work from synthetic output.
Third-order effects
If platforms normalize AI accounts as participants, the social graph shifts from a record of human relationships toward a mixed human-and-machine distribution system, with authenticity and provenance becoming core governance questions.
This is an early test of whether social platforms can turn abundant synthetic supply into engagement without eroding the user trust that makes social content valuable.
The trend: The story belongs to the shift from AI as a user-invoked assistant to AI as an autonomous producer and distributor of platform content.
no further sign of tech executive brain rot re: AI than meta making this announcement like it's something exciting that people actually want. everyone i have talked to this holiday season hates how much slop is on social media. i have barely been on instagram recently bc of it
The future is social media filled with AI-powered bots? At least according to Meta's AI vision? Would you consider it useful to interact with artificial, inauthentic agents? This may be useful, for example threading bots, image generators. www.ft.com/content/9118... [image]
the bull case for meta is that the bots get better and people talk to them the bear case is that the tech just isnt good enough, bots spam the feed and/or people get upset, like when FB leaned harder into discovery by injecting posts in your feed from people you dont follow
sam could be right here if this is in regards to this meta interview this idea is essentially a self-enrichment mechanism for meta platforms to never have the problem of running out of posts the question is if (or when) it starts churning out a never-ending loop of garbage [image…