I think you have an outdated understanding of AI workflow. They generally cite their sources, which you should check, just like regular search.

The source is not what’s convincing you, it’s the way the ai is presenting you the information. The source just confirms what you’re already thinking at that point (what the ai has just presented to you). You’re still trusting the ai.

That may be true of the average user, but you have no way of knowing it's true of the person you replied to. It's 100% possible to check the sources properly, and form your beliefs accordingly, if you want to.

We’re so cooked, all the thinking outsourced to LLMs.

Yes, the AI is trained on a vast quantity of data therefore it is less likely to be manipulated vs a single editor that may have ulterior motives. Therefore it's much harder to manipulate. A corporation which represents many shareholders' interest has its own reputation on the line, which would be seriously damaged if they were caught doing anything like you suggest.

But this can only be understood within the context of the white genocide currently happening in South Africa. Some are saying it's not real, but there have been documented attacks on farms and chants of "kill the boer".

> Yes, the AI is trained on a vast quantity of data therefore it is less likely to be manipulated vs a single editor that may have ulterior motives.

Gemini was caught stuffing prompts with "custom" keywords on certain requests, so there is still an editor between you and the AI.

Read my whole comment :)

I see what you did there :)