most of the time you’ll be talking to a bot there without even realizing. they’re gonna feed you products and ads interwoven into conversations, and the AI can be controlled so its output reflects corporate interests. advertisers are gonna be able to buy access and run campaigns. based on their input, the AI can generate thousands of comments and posts, all to support your corporate agenda.

for example you can set it to hate a public figure and force negative commentary into conversations all over the site. you can set it to praise and recommend your latest product. like when a pharma company has a new pill out, they’ll be able to target self-help subs and flood them with fake anecdotes and user testimony that the new pill solves all your problems and you should check it out.

the only real humans you’ll find there are the shills that run the place, and the poor suckers that fall for the scam.

it’s gonna be a shithole.

  • umbraroze@kbin.social
    link
    fedilink
    arrow-up
    11
    ·
    1 year ago

    SubredditSimulator was based on older generative algorithms, so everyone could fairly easily tell that it was rubbish. When SubredditSimulator got shut down, someone made a new one based on GPT-2 (I think) and everyone was like “OK, this is getting harder to distinguish from real people”.

    I’m betting someone has made even more advanced bots by now. I’m betting someone’s also not concerned about telling other users upfront that they’re bots, and they’re not confining them into specific subs. Now, the only reason I’m not accusing Reddit Inc themselves of building these bots is that they aren’t exactly a bastion of software engineering excellence; the site barely works as is.