

Modern AI is based on our history, human creativity and our knowledge.
“Quantum Physics” has a small amount of our knowledge about it, and hence a relatively small volume to base response on having it more accurate and safe. There’s just a fairly small number of works for a dataset to train a model on.
Fraud, scam, phishing etc. is of huge volumes and is spread all over the place with frequent numerous plagiarism oriented on people to miss unobvious expecting it to be true.
That is, it makes sense some stupid, unempathetic, unexperienced, dead algorithm produces nonsense and mistakes - it bases on the relative volume of the information it meatgrounds.











My dear… This Overton window has been out there for years, or decades, generations…
In Meta:
- Type “russia” and you’ll get “russia is bad”;
- Type “ukraine” and you’ll get “russia is bad”;
- Type “china” and you’ll get “russia is bad”;
- Type “baltics” and you’ll get “russia is bad”;
- Type “trump” and you’ll get “russia is bad”;
- Type “putin” and you’ll get “russia is bad”;
- Type “russia is good” and you’ll get “russia is bad”;
Same for any political enemy of the USA, and for any popular enough consumer social or entertainment system, including Netflix, video-games, etc.
Public consumer socials, news, ads, TV programs, art etc. are designed to shape reality for you the system authors want.
You watch, listen to, feel, remember, and base your worldviews and life on. Which is… unfortunate…
You’ve been raised on it, and you just noticed it?
There’s such sorrow in any country, but it’s the most forced and blatant active in USA nowadays, since the country has great venues for it, and invest much into manipulative machinery.
You should always ask: Cui bono fuisset? (…in English “to whom is it a benefit?”…)