The thing that brings me to solid is my idea of building a personal AI infrastructure.
In this system, the AI agent should serve as a personal data hub, authorization and authentication center. This agent has all the daily data and internet activities the user has. He sees what user sees, learns what user learns, he get personal data when users doing any online activities through it quietly, then it can provide the user with all kinds of personalized digital services, among of them, I can see several most promising areas like in personal financial advice, dating advice( AI talk to AI to help you decide if the owners are compatible), etc.
I was having this idea for a while, find that solid could be a perfect platform to build it so that users trust this AI infra because they control their own data.
This does sound interesting and I love the idea of everyone having a personal AI that works entirely in their interest rather than in the interest of some private organization. I know next to nothing about AI and ML unfortunately so forgive my ignorance but I have a few questions:
What data would be used to train the AIs? Would it be entirely trained on the data of its owner? If so, how would it develop a reference points for comparison with other peoples’ personal AIs? Perhaps there could be some sort of abstraction layer where the users’ data is not exposed but certain statistics about their data are… and that could be used as the basis for the AIs’ interaction with each other?
Okay I guess I only have one question right now lol but it’s an interesting idea and I’m curious about how it would work. I’ve also been curious about “decentralized AI” recently (though I’m not sure how that works either tbh lol) and this seems related to that.