AI is OK, but not when the company only conceives it to obtain user data and does not control where the LLM obtains its knowledge bases, allowing them to indiscriminately scrape the network.
I have been using an AI Search for 2 years (Andisearch), little known, from a small Staetup, that is different from all the others in all aspects, so far the only decent and reliable AI I found.
It has its own LLM, no focused on chats, but to "understand" the concept of your question, giving a direct answer, verified from reliable sources on the web in real time, which is also listed in a search result. 100% private, anonymous, no ads, logs, random proxy, with own reader mode and sandboxed videos.
It don't invent nothing, if it don't know the answer, offers a general web search instead. This is the correct way to use an AI.