A highly sophisticated AI-based aged account sockpuppet swarm, for the purposes of this market, might look like the following:
Someone registers a bunch of accounts on various social media sites/fora (or buys/hacks a bunch of existing ones)
That someone uses AI to create appropriate credible personas for them (ie, if the account's brand is knitting, it mostly sticks to knitting, but it also just so happens to have an interest in biomedical science, or geopolitics, or whatever the actual payload target is, and it credibly establishes this interest in its posting history for quite some time, well before it is "called in" for a strike
Each account has its own unique voice that is driven by AI
Each account has an established timezone its supposed to be posting from, and it sticks to that posting schedule
And then at some point, Mr. Someone puts out the bat signal and all these posters -- who just so happen to have been credible long term posters of special targeted communities, descend on someone who is wrong on the internet.
If we see credible proof of something that mostly* pattern-matches to this, and it happens before the end of 2025, this market resolves as YES.
*It can be missing a characteristic or two (ie, they goof and giveaway their timezone, or they never call it in for a specific strike, they just use it "build consensus" in their direction constantly through background chatter), but the absolutely essential characteristics are the aged account aspect, and that each one is fairly differentiated/customized to its persona, and not just all blasting out with the same voice.
1,000People are also trading
The market description frames it as:
If we see credible proof of something that mostly* pattern-matches to this, and it happens before the end of 2025
I take that to mean that this market is not just asking if these "highly sophisticated AI-based aged account sockpuppet swarms" will exist before the end of 2025, but that we'll see credible proof for it. Indeed, that's the way the question itself is worded—"Will we see highly sophisticated AI-based aged account sockpuppet swarms before the end of 2025".
It makes sense for the question to be set up that way, otherwise it could never resolve NO! How can you prove that there _aren't_ such swarms that are so so highly sophisticated that they have evaded all possible detection, and will continue to avoid detection indefinitely?
I think that unless someone can provide here the "credible proof" requested in the description, this market should resolve NO.
@zzq Yes, my guess is that they do already exist and are being used in practice by nation-states at the very least, but this market should probably resolve NO given the lack of definitive evidence.
The freedom house reports linked in /firstuserhere/will-we-find-out-in-2023-about-a-na have a lot of usage of AI-sockpuppet aged accounts being used.
.
Evidence of (failing) swarms in action: https://twitter.com/conspirator0/status/1647671394476478467?s=20
As a real human who decided to post about stuff beyond my "brand" I kind of almost take offense at this being considered botlike behavior. "Bot" is already a derogatory and literally dehumanizing way to refer to human accounts (and even IRL folks with the NPC term.) Bots/deepfakes/catfishers/etc are a real and growing problem, but I also really don't like the direction things are going from the perspective of the response to this problem.
@cloudprism This market isn't saying venturing outside your area of established interest is botlike behaviour - on the contrary, it's so human-like that until recently it has all-but-guaranteed the poster is a real human. It's because of this credibility that we're now worried that a high-effort bad actor could abuse this precedent to invisibly wield an army of bots to shift opinion within a subculture.
@jacksonpolack the tech is definitely here, it's just a matter of execution , eg https://twitter.com/blader/status/1642316995004882944?t=utsCS1ScizTlyNvguv-HsA&s=19