Welcome to the Misleading AI Myths

Please look around you and you will realize that nearly everyone talks about AI. It does not matter whether it is true or fiction, everyone has some ideas and thoughts to share. It is an undeniable fact that Hollywood has been producing films related to technology and artificial intelligence and this genre is populating the big screens. Moreover, in today’s world, it is easy to reach those movies or series out. However, whatever the main idea is, robot armies or super intelligent computers had created fears and uncertainties about the future and maybe that is why some myths come from this origin.

I started to read a book called Being human in the age of Artificial Intelligence, by Max Tegmark. The author mentions how important is the rise of artificial intelligence and machine learning over the last few years and makes several evaluations. The chapter of myths grabbed my interest and I would like to share my thoughts with you.

Myth 1: According to the book, one popular myth is that we will get superhuman AGI this century whereas a counter-myth is that we will not get superhuman AGI in this century. What is AGI? Artificial General Intelligence means the ability to accomplish any cognitive task at least as well as humans. In my opinion, we cannot say anything about both with 100% for sure because even though our history and current century are full of technological developments, we still do not have flying cars or power plants. On the other hand, according to estimations, we cannot say with high confidence we are far from superhuman AGI.

Myth 2: There is a legitimate fear of AI taking away our jobs soon. Actually, this point is not about losing our jobs; it is about a structural change of roles and positions. For example, there are many repetitive jobs such as Customer Service over the business world, and here AI can be trained to automate the process and replace a human mind taking care of it. In this case, people will be challenged to learn and work on different tasks and also improve their skills in new areas in which human skills are sensitive. Even displacement is a foregone conclusion between people if there are more educated or qualified people, why we do not try to use the advantage of AI in a good way?

Myth 3: AI will turn evil and smarter than people. There is a point that needs to be reminded; AI can be as smart as we program it and can be turned competent with goals. If there are no humans, there will not be artificial intelligence. For sure, we can train them to think faster than us, but it does not mean that the decisions will be always right. Even having the smartest engine known: the human brain, we are still falling in a lot of mistakes or bad reasoning while making decisions, so how could be AI smarter than us?

Myth 4: I am going to tell you something that you already knew or heard about. Yes, data! We need big data to train AI. To get an accurate result, we need to train and test data, and if we think we have data lake, the results would be with an amazing accuracy rate. However, we might be missing some important parts. We do not need only big data; we also have to make sure this is good quality data since we need to invest a considerable effort into analyzing in different areas such as data analysis, feature modeling, and model selection before implementing algorithms. Shortly, there are also many statistical steps behind the processes. At the very end, Machine Learning algorithms are trained with large amounts of data (%80) and tested with less data (20%). Therefore, instead of big data, AI systems should start with simple tasks, and then they will be able to learn and increase capabilities to get outcomes that are more accurate.

Myth 5: Machines cannot have goals. Honestly, this is completely wrong. As I mentioned above, we train and program them, which is why if their goals are aligned to ours, they cannot be wrong. For example, a heat-seeking missile has a goal.





Leave a Reply