3 Myths about Artificial Intelligence, the Human Brain and Autonomous Cars: Why we will not see a real autonomous car in the foreseeable future

Knowing how your brain functions will help you to understand, why even IT professionals, grossly overestimate our wonderful artificial intelligence (AI) and underestimate our flabbergasting human brain. As a result, even very clever and successful executives, IT-directors, managers, tech entrepreneurs and tech professionals throw away massive amounts of time and money to realize ideas that are nothing but pipe-dreams.

This lack of knowledge about the human brain also makes other people needlessly afraid of a superior “general artificial intelligence” that in a near future will surpass human intelligence, while in reality the greatest present threat is the artificial stupidity of our present-day “narrow artificial intelligence”.

This book compares the abilities and limits of a human brain with artificial intelligence. To make the differences much easier to understand the author uses the example of a human being driving a car and an autonomous car.

AI does not replicate general human intelligence, it does not even scratch the surface. It only imitates a few narrow aspects of the way the human reflex-brain-system functions, while attempting to bypass the meaning of the data it processes. This limited imitation of the functioning of the human reflex-brain-system is sufficient to do routine tasks, even very complex ones, that can be reduced to statistics, to mathematical models, to a (massive) computing of probabilities. It often performs these components amazingly better and faster than a human brain, although at an extremely high energetic cost.

Since computer systems can do this much better and faster than people, there is no doubt that they will be able to drive a robot-car in a simplified, predictable environment, adapted to their limitations and with a very fast broadband connection to tap a database of information to deal with rare events.

AI, however, will not be able to do what our think-brain does to drive a car in unpredictable, unforeseen, chaotic, new and unique situations, where other road users are unpredictable, and don’t always follow the rules. To do that we need AI systems that can process abstract concepts, representations of reality, or at least that can imitate the processing of meaning, ideas and common sense.

Until then, when its databases do not have an answer, the robot-car will have to learn to execute a safe stop in all circumstances, and call for the help of a human brain. This will be difficult to do if the decision needs to be made in a split second, like our think-brain does.

Finally, the greatest threat today is not artificial intelligence at all. It’s even not the ever receding fata morgana of general artificial intelligence. The greatest danger is the present “artificial stupidity” of our narrow artificial intelligence grossly amplifying the human lack of knowledge, stupidity or wickedness.