Most of us know by now that many popular online forums are plagued with bots, some more advanced than others.
I propose that they are not only participating in existing debates but creating entirely new debates and re-kindling old ones, such as with the flat earth idiocy.
What if the goal of alphabet agency X is to test its advanced AI to see how well it can influence online debate, change opinions, and get people to waste their time. starting from a completely ridiculous foundation such as 'flat earth'?
view the rest of the comments →
Wargasm ago
There are easier, and less retarded, ways to test advanced AI.
For instance, wouldn't it be a more useful test to see how a covert AI interacted with real people on social media? That test could demonstrate worth in fields like customer service with actual numbers like: number of human friends made, number of actual human interactions without detection, etc...
lemon11 ago
That sounds like exactly what a FE bot already would do. Influencing dumb conversations reduces the costs and risks of exposure, which means they can operate longer and more flexibly to perfect the technology.