That's more than a slight nitpick - it corrects a core misunderstanding to be found everywhere this discussion seems to be happening. To wit: AI articles are t

That's more than a slight nitpick - it corrects a core misunderstanding to be found everywhere this discussion seems to be happening. To wit: AI articles are to be found along a dramatic spectrum of quality. If such an article is high-quality, relevant to the subject matter, asserts the fact in question, and makes proper and veracious use of a primary source in support of such an assertion, why isn't it a reasonable source for an encyclopedia? I can imagine a future with rich educational materials with these layers: * raw experimental data -=> * publication ("primary" source) -=> * AI-generated review of many publications ("secondary" source) -=> * Human-authored encyclopedia article (with one or two people following the fact pattern all the way back down to the data, and many more people helping to synthesize the higher layers into a rich, readable, considerate, diverse synopsis)