The researchers’ bots generated identities as a sexual assault survivor, a trauma counselor, and a Black man opposed to Black Lives Matter.
The researchers’ bots generated identities as a sexual assault survivor, a trauma counselor, and a Black man opposed to Black Lives Matter.
The issue is, with AI you can have an agent personalized for every person, pretending to be their friend, manipulating them individually, to move society in a way that’s negative for 99.999% of people.
Maybe the real friends were the AI bots we made along the way.
I think that’s hyperbolic (at least at today’s “AI” capabilities) all of the posts and comment in this ”research” were reviewed and posted by a human researcher.
Even if the technology improves to the point where that is no longer necessary. Even the billionaires are finding out you only need a few bad eggs to spoil the basket
Really? I can make an agent that can convince me it’s my friend?
You can have an agent that simultaneously has “personal” conversations with a ton of people.