Why are you so sure that typical AI can surpass human intelligence?

This is a bit of a rant but I hear people from the machine learning side always comment on how soon typical AI is going to surpass human intelligence.<br><br>I think this is a huge misconception. <br><br>There are a few different reasons why we are not close to surpassing human intelligence with typical AI.<br><br>**1. Intelligence is a multifaceted phenomenon.** <br><br>Intelligence can sometimes even be completely beyond typical understandings of intelligence. For example a more emotional and empathetic AI would be more intelligent in that respect.<br><br>**2. There are many ways to be intelligent.** <br><br>There are many different forms of intelligence. Spatial intelligence, artistic intelligence, linguistic intelligence. Each of these forms of intelligence could have a completely different path to AI. There is evidence to suggest that certain animals have an understanding of spatial awareness that far surpasses humans. There could be a ton of intelligence in the world that we have no idea about. <br><br>**3. It is impossible to understand intelligence without typical AI.**<br><br>It is impossible to understand intelligence without typical AI because so much of the research on typical AI is vastly oversimplified and uses essentially the same algorithms that have barely changed over the past few decades. We have no idea what intelligence is because we have no idea how typical AI works. <br><br>**4. Typical AI is a vast simplification of human intelligence.** <br><br>Our typical AI algorithms are vastly simplistic representations of the typical synapses, neurons and brain matter that make up the human brain. The human brain is such a complex phenomenon that we have little to no understanding of how it works.<br><br>**5. We have no idea how to even model human intelligence.** <br><br>We have no idea of how to even model human intelligence. We have no idea what intelligence even IS, so we have no idea what we are even working towards. <br><br>I think that the AI research community should focus more on making AI systems that are more biological typical, and less on making AI systems that are so simplistically designed.