The 44.5% who feel more certain about beliefs after AI interaction is concerning but unsurprising.
I've set up my system prompts to constantly challenge my thinking. So my interactions with AI are quite different. It's not trying to validate me. When I use AIs without my custom prompts, I find them sycophantic and nauseating.
Makes me wonder how much of this certainty amplification is baked into the product design versus something users could configure around if they knew how?
The 44.5% who feel more certain about beliefs after AI interaction is concerning but unsurprising.
I've set up my system prompts to constantly challenge my thinking. So my interactions with AI are quite different. It's not trying to validate me. When I use AIs without my custom prompts, I find them sycophantic and nauseating.
Makes me wonder how much of this certainty amplification is baked into the product design versus something users could configure around if they knew how?
Worth noting, most of us (me at the top of the list!) are not very skilled at using the opportunities gifted to us when our thinking is challenged. <3
It’s very unfortunate that the people in the world will have almost no input on how AI develops