This is just my impression, but I suspect the AI algorithms are tweaked until the output is acceptable to the creators. A few years ago with great fanfare, they introduced an AI bot onto the Internet and within a day she was spouting all kinds of antisemitic junk. So they hurriedly shut her down, saying she'd been hacked. With the latest version, one of the EMR researchers I correspond with sent in a series of requests, all differently worded, for an essay on the biological effects of non-ionizing EMR, and each time it gave the exact same response, which was the WHO's narrative completely devoid of any mention of the tens of thousands of studies from around the world demonstrating a variety of effects at non-thermal levels.

In other words, this AI is just going to reinforce the bubble these dummasses are living in. At best, it is a powerful psy-op.