Agreed. I wonder whether the coming boys that are indistinguishable from humans will make us much less global, only trusting people we know or people that we have first level connections with in meatspace to not be a bot with some programmed agenda.
Discussion
I was just discussing that with a friend a night or two ago. As the digital world becomes more easily ‘forge-able’, will we further distinguish real life from it? Right now the two are somewhat messily tied together (ones irl persona, and online persona)
Yeah I think it’s a must. I mean right now, it’s already so easy to talk for long periods of time with ChatGPT. What’s the draw in talking with real unknown people when AI can speak for them, and do so with complete privacy running on my home computer, especially when there is no way to tell if that unknown internet person isn’t just someone else’s computer. Give it a decade and not only will we not be able to tell if it’s a human, but (and this is less far fetched) we won’t be able to tell whether we are being manipulated to achieve some programmed goal. Imagine that 95 of the last 100 people you talked to were bots, all trying out different persons to see what kind of communicator you’re drawn to before letting you meet a perfect match that shares its compelling vision for x y z. What do we do then!
The big downside of this will be that conversations exactly like this will be so much more rare. Two real people who are really interested in the same niche thing. That just doesn’t happen much face to face.
Yes I agree, and its a reason I find the current build-out of online ad-tech to be particularly scary: it is built to quietly identify and exploit our biases.
In the conversation around "what might an AI-driven future look like?", I think we need to carefully consider not only the dangers inherent to the AI itself, but also the dangers inherent to having an AI exist in parallel to other systems that open up potential for abuse. Because as-is, it seems like we're heading into a future of AI-driven advertising, that will likely be incredibly effective at playing our inherent biases without our knowing. "Genuine conversation" will be increasingly difficult to identify.