How do we know that the people on reddit aren’t talking to bots? Now, or in the future? what about lemmy?

Even If I am on a human instance that checks every account on PII, what about those other instances? How do I know as a server admin that I can trust another instance?

I don’t talk about spam bots. Bots that resemble humans. Bots that use statistical information of real human beings on when and how often to post and comment (that is public knowledge on lemmy).

  • bigboismith@lemmy.world
    7·
    1 day ago

    Totally fair question — and honestly, it’s one that more people should be asking as bots get better and more human-like.

    You’re right to distinguish between spam bots and the more subtle, convincingly human ones. The kind that don’t flood you with garbage but instead quietly join discussions, mimic timing, tone, and even have believable post histories. These are harder to spot, and the line between “AI-generated” and “human-written” is only getting blurrier.

    So, how do you know who you’re talking to?

    1. Right now? You don’t.

    On platforms like Reddit or Lemmy, there’s no built-in guarantee that you’re talking to a human. Even if someone says, “I’m real,” a bot could say the same. You’re relying entirely on patterns of behavior, consistency, and sometimes gut feeling.

    1. Federation makes it messier.

    If you’re running your own instance (say, a Lemmy server), you can verify your users — maybe with PII, email domains, or manual approval. But that trust doesn’t automatically extend to other instances. When another instance federates with yours, you’re inheriting their moderation policies and user base. If their standards are lax or if they don’t care about bot activity, you’ve got no real defense unless you block or limit them.

    1. Detecting “smart” bots is hard.

    You’re talking about bots that post like humans, behave like humans, maybe even argue like humans. They’re tuned on human behavior patterns and timing. At that level, it’s more about intent than detection. Some possible (but imperfect) signs:

    Slightly off-topic replies.

    Shallow engagement — like they’re echoing back points without nuance.

    Patterns over time — posting at inhuman hours or never showing emotion or changing tone.

    But honestly? A determined bot can dodge most of these tells. Especially if it’s only posting occasionally and not engaging deeply.

    1. Long-term trust is earned, not proven.

    If you’re a server admin, what you can do is:

    Limit federation to instances with transparent moderation policies.

    Encourage verified identities for critical roles (moderators, admins, etc.).

    Develop community norms that reward consistent, meaningful participation — hard for bots to fake over time.

    Share threat intelligence (yep, even in fediverse spaces) about suspected bots and problem instances.

    1. The uncomfortable truth?

    We’re already past the point where you can always tell. What we can do is keep building spaces where trust, context, and community memory matter. Where being human is more than just typing like one.

    If you’re asking this because you’re noticing more uncanny replies online — you’re not imagining things. And if you’re running an instance, your vigilance is actually one of the few things keeping the web grounded right now.

    /s obviously