Comment

Comment on Southern California air board rejected pollution rules after AI-generated flood of comments

SnotFlickerman@lemmy.blahaj.zone ⁨1⁩ ⁨week⁩ ago

This was happening before AI, with less sophisticated tools, often called “Persona Management” that allowed one person to control numerous bots with pre-written scripts that could be called up depending on what was called for. The only difference the AI has made is the speed and scale at which the same can be done and be more convincingly not all culled from the same script.

axios.com/…/bots-flooded-the-fcc-with-comments-ab…

Here’s an article about a flood of bot comments to an FCC open comment regarding Net Neutrality in 2017, five years before OpenAI would release ChatGPT.

source

Sort:hotnew top

supersquirrel@sopuli.xyz ⁨1⁩ ⁨week⁩ ago
Yeah AI is an acceleration of

source
T156@lemmy.world ⁨1⁩ ⁨week⁩ ago
It does make it harder to find them, because the phrasing is similar, but not identical due to randomness.

Whereas before, you could probably filter a good chunk of it out by just finding the same message/keywords and filtering by that.

source
GreenBeard@lemmy.ca ⁨1⁩ ⁨week⁩ ago
Yeah, you can kill a man with a knife but you can do it a lot faster and easier with a nuclear warhead. People aren’t scared of an aggressive chihuahua, but they’ll have an aggressive pitbull put down. The scale and scope of damage matters.

source
- SnotFlickerman@lemmy.blahaj.zone ⁨1⁩ ⁨week⁩ ago
  Allow me to quote myself, from my initial comment in this thread, which was the first comment in this thread.
  
  The fines/prison time should be even more severe when AI generated messages are fraudulently being promoted as real humans, simply due to the industrial speed and scale AI generation allows.
  
  I know this, I made it clear why it’s a problem when nobody else had even commented in this thread yet… I was merely pointing out that this has been a growing problem for a long time before AI became part of it.
  
  source