However, it gave me a ton of exposure to the 5th generation project, and this article is a wonderful reminder of how crazy it was, though it doesn't quite capture the insane hype around the whole thing.
There's an excellent, if imperfect, book called "The Fifth Generation Fallacy" by JM Unger (who, in a weird circumstance that makes me think the writers of our narrative are just plain lazy, a few years later, became the head of the Japanese department I was in) that talks about the project in terms that I think make quite a lot of sense.
The book talks about it, but I go even further: the 5th gen project was about Japanese beliefs about information technology and specifically that Japanese was such a complex language that the only way to have working IT was to have an intelligent machine.
Japan was still using pencils and xerox machines in the aftermath of the PC revolution in the US and Europe (in practice, this continued into the mid-90s), and had missed mainframes and minis entirely. The idea of a statistically-driven, convenient, modern input mechanism (what people use now when they use an IME, these started to show up in the mid-90s when enough of a corpus was analyzed to make them work) was not there yet; Japanese word processors of that era are incredibly cumbersome and terrible, just barely usable and not usable at all by large swaths of the work force. Japan was absolutely terrified of the technological progress in the rest of the world and a lot of it had (and has) to do with the simple inconvenience of the writing systems and the linguistic isolation.
In the cultural context, especially of the bubble that formed at the same time, and no small amount of vaguely racist beliefs about their superiority (especially linguistic, which had been going on for a very long time and played a role in Japanese relations and eventual behavior post-annexation of Korea), the 5th gen project makes perfect sense.