I have been following the "woke AI" stuff for a few days. I thought it would all blow over, but then it kept getting worse (and funnier). We've all known for several years that Google has manipulated search results, and really if you want to find something controversial in some way, you should use DuckDuckGo or Brave.
Anyway, the AI has been programmed for diversity, equity, and inclusion so naturally it provided images of "nonwhite Nazis", black Founding Fathers, black kings of England, etc, it defended pedophilia and referred to pedos as MAPs, when asked to compare conservatives to Hitler it would tell you that it was complicated and the conservative might be as bad as Hitler (with more words, obv). I know it has made up a conviction record for a woman that I follow.
https://www.nbcnews.com/tech/tech-news/google-says-gemini-ai-glitches-product-effort-address-traps-rcna140243https://www.foxnews.com/media/red-faced-google-apologizes-woke-ai-bot-gives-appalling-answers-pedophilia-stalinGoogle is spinning this as "glitches" but as soon as a query goes viral on Twitter, Google shuts it down (the AI stops answering that particular question).