I tend to use a lot of services from Microsoft, Google, and Apple to do my personal research, write my books, and for occasional consulting projects. There are a lot of trade offs and decisions people need to make for themselves.

Two books are useful: Surveillance Capitalism, and The Power of Privacy.

Not as much privacy oriented, but I find the Center for Humane Technology https://www.humanetech.com/ organization a good source of thoughtful material.

Off topic, but after writing my last book on using LangChain/OpenAI/Hugging Face with your own data (you can read it online for free https://leanpub.com/langchain/read) I am looking at AI and personal knowledge management from a perspective of maintaining personal control and privacy in a new book I just started writing Safe For Humans AI (and grabbed the domain with the same name). This is a difficult problem unless you either own very high end compute at home or are willing to rack up major cloud spend. Using top-K text chunk matching using embeddings (LangChain and LlamaIndex) is much cheaper and easier than fine tuning public models, but both techniques have their own applications.