Before looking for answers, we should understand intersubjective realities first. An apple is an objective reality. It exists regardless of what we think about it. Love is a subjective reality. It exists within our mind. There’s a third category: intersubjective reality. Countries, money, companies belong to this category. They exist because billions of people collectively agree that they exist. If people stop believing in them, they can’t exist. Countries collapse, companies fail, and currencies lose value when they loose credibility. There’s another popular intersubjective entity that people forget overlook. Humanism.

People tend to think that human beings are rational individuals with logical free will. No, we are not. We are more stupid than we think, we make surprisingly random decisions, and free will is actually a fiction. That is why it’s so easy for algorithms to manipulate us. People consume what algorithms recommend and vote for parties that algorithms expose. AI is often better when it comes to making logical decisions. If that is true, how can we keep insisting humans should be the ones who vote, consume, and rule the world?

Another common misunderstanding is the belief that AI is infallible. AI works 24 hours, knows vast amounts of knowledge, and provides the most effective answer. We already know something that had the same power: God. People easily make mistakes considering AI as a new god. They ask everything to AI, receive answers, and never doubt its holly answers. However, there’s so many evidence that AI makes countless errors and carries significant biases. Even worse, it’s often difficult to recognize when AI is wrong due to its incomprehensibility. That’s why AI will always require an intentional bottleneck: a human supervisor.