Twitch, one of the top live-streaming platforms active today, had a streamer suggest based on audience interaction figures that somewhere around 80% of the top 500 streamers were artificially inflating their viewcounts with bot viewers. Archive of our Own, a popular fanfiction website, is struggling to strike a balance between requiring an account, which reduces interaction, and allowing anonymous AI reviews, seeing as the current surge of AI reviews are often incredibly critical and negative (the exact logic to this is yet to be puzzled out). Spotify has been accused of using AI to generate muzak, as playlist filler, but bots are also accused of artificially boosting the ‘listens’ there, and X, BlueSky, Instagram, et cetera are rife with bots trying to sell things, trying to scam people, trying to boost interaction numbers, flaming, and more.
So, the bots are everywhere. The bots are getting better and better at pretending to be human with generative AI on their side, to the point that real people are effectively failing the Turing test even when they ‘know’ their conversation partner is a bot. The bots are counted as people by most view metrics on most view-powered sites.
Why are we doing this?
Each individual system is benefited by AI for their particular scheme. Advertisers hate bot viewers because those cost them views but are not potential sales, but the bot viewers boost the revenue of the streamers and Instagram influencers. Conversely, relatively unmoderated comment sections are perfect spots for bot commentors to leave links in the hopes someone will click it, something the streamers et al. hate! The real Spotify artists hate the bot music, as the bot music means Spotify doesn’t have to pay a real artist for streams, and Spotify hates the bot listeners because it makes them pay out anyway.
While each individual benefits, the overall result is worse for everyone. Small artists are not sure how many people are actually listening or watching their content, and how many of those will convert to sales, large platforms are disincentivized from paying based on views, with no really good alternatives, and would-be commentors are lost in a flood of spam messages.
Even in instances where the bot is not trying to trick others into believing they’re a person, the result is still worse for everyone than real interactions would be. Companion chatbots can earn a couple of bucks off people who are desperately lonely, and while they’re not trying to steal these people away from their loved ones, they’re so convincing that it’s almost impossible for someone in that state of mind to resist them. After all, the people who are well-adjusted enough not to need an AI companion may be the target audience, but they’re not the actual audience. The actual audience are frequently people struggling socially, who may or may not have a mental illness that the bot will exacerbate into a full-blown detachment from reality.
There’s friction to fixing these problems in every direction. While view bots are bad for big business, they’re tough to pin down and weed out. Where AI-generated muzak is annoying to human artists, there’s a monetary incentive to keep it on the platform. Unless everyone can agree it all has to leave, there will be no real answer to the problem.
Worried about DDoS attacks, or protecting your website from bots? Contact us! https://elixistechnology.com/contact/
Sources: https://www.dexerto.com/twitch/devin-nash-reveals-shocking-viewbotting-details-about-twitchs-top-streamers-3232885/
