Why artificial intelligence is not intelligent at all and is only called that for marketing reasons.

Admittedly, today's AI models not only consume a lot of electricity but also deliver results that surprised me, but also absurd errors that are trivialized as hallucinations. One fact that is often overlooked, partly due to good marketing, is that today's AI models are based on calculated probabilities. The AI does not really understand what it is doing, it only calculates probabilities. Admittedly, it often does this very well until a “hallucination” occurs. This means that AI can only reflect human knowledge, human perspective and human bias and racism. Everything is based on human training data and the AI calculates with which probability the human user would like to have which knowledge. Wouldn't you expect much more from real intelligence? I just wonder how people can attribute intelligence to this well-calculating machine? There are people who even fall in love with today's AI models or see them as real artificial life. Fueled by marketing that promises us a real super intelligence in the near future in order to get even more venture capital to maintain the whole hoax, which devours vast sums of money in data centers and electricity. Another more serious problem in my view is that today's AI models are destroying the jobs that we need to continue creating good training material for AIs. Today's AIs can easily replace graphic designers, translators, authors, musicians and poets. As a result, the internet is already flooded with AI-generated content. This will eventually become the new training data, which will further reduce so-called intelligence. How a superintelligence will ever emerge from this is a mystery to me.

#ai #technology #training #internet #user #marketing #economy #money #business #future #humanity #problem #ethics #philosophy #fail #knowledge #science #intelligence