Thread by @elisethoma5

1. I've found a large, AI-driven bot network on X which is posting in support of Trump in the US election.

Thread.

This network seems to be big but crude. The accounts are an eclectic mix of join dates, handles and personas which are mostly* not tailored to pro-Trump activity. This is pretty typical of crypto or spam networks, which tend to be cobbled together out of accounts like this.

Many are blue tick verified, which again is super standard for spam networks.

* Some of them ARE tailored, with some version of 'Trump supporter #MAGA #2A #KAG' whatever whatever in their bios, even when the earlier content is completely incongruous with that persona.

A small number of accounts have more elaborate personas. These mostly appear to have been active since late June 2024 and act as central nodes in the network. I'd call these originators and the others amplifiers.

Why am I so confident that these are AI bots and not real people? Well for one thing, they told me so.

Occasionally they post refusals, although it happens rarely enough that whoever built the network has clearly found a way around @OpenAI's safeguards which works fairly well.

Sometimes they get confused by the prompt. Here, for example, an originator posts 'Distinguished' and the amplifiers are mystified.

Sometimes, delightfully, they also argue WITH THEMSELVES.

I have no idea what quirk of AI is making that happen but it's very funny.

If you're wondering what the poems have to do with anything - so am I. They appear to have a small rotation of random images they post alongside their text regardless of the disconnect between the two. This again is bog standard bot behaviour.

Interestingly many of them are using the hashtag #Trump2020. Little late, guys?

Outside of their own network, it looks like (on a manual review because X doesn't let us do the fancy stuff any more) the account which they seem to tweet at the most actually isn't Donald Trump, it's Elon Musk. E.g. x.com/search?q=from%…

Sidebar but this is an interesting post. I'm guessing this is some sort of guardrail within @OpenAI kicking in and preventing the bot from endorsing Musk's election fraud nonsense.

Before switching to the pro-Trump activity, many of the accounts appear to have been used for fairly standard spammy stuff - porn, cute cats, sports. These appear to have been largely commercial bot accounts.

There is no indication that I've found yet which suggests who might be behind this network (although if I had to take a wild swing at geography, I'd be thinking Pakistan or India).

The gamechanging thing about AI is that networks like this could be pretty much entirely automated. I don't think anyone is even reading these tweets before they go out, or else the many many mistakes would have been caught. This could be the work of...

a group, but it could also easily be the work of just one person. As I wrote for @lawfare recently, it's really important not to jump to conclusions about attribution and ESPECIALLY so now we're in the GenAI era lawfaremedia.org/article/genera…

@lawfare At this stage it doesn't look to me like the network is generating much authentic engagement or likely to be influencing any real person's opinions. It's interesting but not a five alarm fire.

@lawfare And now I have to go and get my laptop fixed (do you have any idea how annoying it is to type with a broken E key? Do you?!) but I might have more of a look into this later

Au revoir DJ Trump Nation and Trump Right All.

And au revoir to most of their friends. This platform CAN do content moderation... when they decide to.

I'm seeing some speculation in the comments that this was a Russian state-linked thing. To be clear, there's no indication of that (and I personally don't believe it).

I did a whole thread just last week on not assuming state actors are behind everything