At least with Wikipedia the manipulation is visible and attributable. These LLMs are black boxes to the outside world, rife with manipulations (as can be seen from the way ChatGPT has changed its behaviour over time in specific areas) and in no way verifiable. Even if the creators vow to having used a certain set of data to train the model this can not be verified.

To the downvoters: explain, don't just knee-jerk your disapproval. I don't care about karma but it is tiring to see things which don't follow the narrative being downvoted out of existence.