I've been researching AI's psychological and sociological effects on individuals and humans as a whole. It's part of a broader effort of mine to get my own harness (AGENTS.md, Skills, agent defs etc.) to remain a tool I use and not a tool that uses me.

My thinking is if the AI is directed to no longer provide you with anything other than what you've specifically asked for, then maybe that will limit the amount of influence it has on the human decision-making context.

I think AI is pretty great but there are a ton of problems with it. I'm hoping that this gets out there widely enough to maybe give me some additional data points, but also hopefully to encourage more people to take the time to think and reason through things that they should care about.

I'm interested in hearing nuanced views on the topic.